ARM queue: mostly patches from me, but also the Smartfusion2 board. thanks -- PMM The following changes since commit 9ee660e7c138595224b65ddc1c5712549f0a278c: Merge remote-tracking branch 'remotes/yongbok/tags/mips-20170921' into staging (2017-09-21 14:40:32 +0100) are available in the git repository at: git://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20170921 for you to fetch changes up to 6d262dcb7d108eda93813574c2061398084dc795: msf2: Add Emcraft's Smartfusion2 SOM kit (2017-09-21 16:36:56 +0100) ---------------------------------------------------------------- target-arm queue: * more preparatory work for v8M support * convert some omap devices away from old_mmio * remove out of date ARM ARM section references in comments * add the Smartfusion2 board ---------------------------------------------------------------- Peter Maydell (26): target/arm: Implement MSR/MRS access to NS banked registers nvic: Add banked exception states nvic: Add cached vectpending_is_s_banked state nvic: Add cached vectpending_prio state nvic: Implement AIRCR changes for v8M nvic: Make ICSR.RETTOBASE handle banked exceptions nvic: Implement NVIC_ITNS<n> registers nvic: Handle banked exceptions in nvic_recompute_state() nvic: Make set_pending and clear_pending take a secure parameter nvic: Make SHPR registers banked nvic: Compare group priority for escalation to HF nvic: In escalation to HardFault, support HF not being priority -1 nvic: Implement v8M changes to fixed priority exceptions nvic: Disable the non-secure HardFault if AIRCR.BFHFNMINS is clear nvic: Handle v8M changes in nvic_exec_prio() target/arm: Handle banking in negative-execution-priority check in cpu_mmu_index() nvic: Make ICSR banked for v8M nvic: Make SHCSR banked for v8M nvic: Support banked exceptions in acknowledge and complete target/arm: Remove out of date ARM ARM section references in A64 decoder hw/arm/palm.c: Don't use old_mmio for static_ops hw/gpio/omap_gpio.c: Don't use old_mmio hw/timer/omap_synctimer.c: Don't use old_mmio hw/timer/omap_gptimer: Don't use old_mmio hw/i2c/omap_i2c.c: Don't use old_mmio hw/arm/omap2.c: Don't use old_mmio Subbaraya Sundeep (5): msf2: Add Smartfusion2 System timer msf2: Microsemi Smartfusion2 System Register block msf2: Add Smartfusion2 SPI controller msf2: Add Smartfusion2 SoC msf2: Add Emcraft's Smartfusion2 SOM kit hw/arm/Makefile.objs | 1 + hw/misc/Makefile.objs | 1 + hw/ssi/Makefile.objs | 1 + hw/timer/Makefile.objs | 1 + include/hw/arm/msf2-soc.h | 67 +++ include/hw/intc/armv7m_nvic.h | 33 +- include/hw/misc/msf2-sysreg.h | 77 ++++ include/hw/ssi/mss-spi.h | 58 +++ include/hw/timer/mss-timer.h | 64 +++ target/arm/cpu.h | 62 ++- hw/arm/msf2-soc.c | 238 +++++++++++ hw/arm/msf2-som.c | 105 +++++ hw/arm/omap2.c | 49 ++- hw/arm/palm.c | 30 +- hw/gpio/omap_gpio.c | 26 +- hw/i2c/omap_i2c.c | 44 +- hw/intc/armv7m_nvic.c | 913 ++++++++++++++++++++++++++++++++++------ hw/misc/msf2-sysreg.c | 160 +++++++ hw/ssi/mss-spi.c | 404 ++++++++++++++++++ hw/timer/mss-timer.c | 289 +++++++++++++ hw/timer/omap_gptimer.c | 49 ++- hw/timer/omap_synctimer.c | 35 +- target/arm/cpu.c | 7 + target/arm/helper.c | 142 ++++++- target/arm/translate-a64.c | 227 +++++----- default-configs/arm-softmmu.mak | 1 + hw/intc/trace-events | 13 +- hw/misc/trace-events | 5 + 28 files changed, 2735 insertions(+), 367 deletions(-) create mode 100644 include/hw/arm/msf2-soc.h create mode 100644 include/hw/misc/msf2-sysreg.h create mode 100644 include/hw/ssi/mss-spi.h create mode 100644 include/hw/timer/mss-timer.h create mode 100644 hw/arm/msf2-soc.c create mode 100644 hw/arm/msf2-som.c create mode 100644 hw/misc/msf2-sysreg.c create mode 100644 hw/ssi/mss-spi.c create mode 100644 hw/timer/mss-timer.c In v8M the MSR and MRS instructions have extra register value encodings to allow secure code to access the non-secure banked version of various special registers. (We don't implement the MSPLIM_NS or PSPLIM_NS aliases, because we don't currently implement the stack limit registers at all.) Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-2-git-send-email-peter.maydell@linaro.org --- target/arm/helper.c | 110 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 110 insertions(+) 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 @@ uint32_t HELPER(v7m_mrs)(CPUARMState *env, uint32_t reg) break; case 20: /* CONTROL */ return env->v7m.control[env->v7m.secure]; + case 0x94: /* CONTROL_NS */ + /* We have to handle this here because unprivileged Secure code + * can read the NS CONTROL register. + */ + if (!env->v7m.secure) { + return 0; + } + return env->v7m.control[M_REG_NS]; } if (el == 0) { return 0; /* unprivileged reads others as zero */ } + if (arm_feature(env, ARM_FEATURE_M_SECURITY)) { + switch (reg) { + case 0x88: /* MSP_NS */ + if (!env->v7m.secure) { + return 0; + } + return env->v7m.other_ss_msp; + case 0x89: /* PSP_NS */ + if (!env->v7m.secure) { + return 0; + } + return env->v7m.other_ss_psp; + case 0x90: /* PRIMASK_NS */ + if (!env->v7m.secure) { + return 0; + } + return env->v7m.primask[M_REG_NS]; + case 0x91: /* BASEPRI_NS */ + if (!env->v7m.secure) { + return 0; + } + return env->v7m.basepri[M_REG_NS]; + case 0x93: /* FAULTMASK_NS */ + if (!env->v7m.secure) { + return 0; + } + return env->v7m.faultmask[M_REG_NS]; + case 0x98: /* SP_NS */ + { + /* This gives the non-secure SP selected based on whether we're + * currently in handler mode or not, using the NS CONTROL.SPSEL. + */ + bool spsel = env->v7m.control[M_REG_NS] & R_V7M_CONTROL_SPSEL_MASK; + + if (!env->v7m.secure) { + return 0; + } + if (!arm_v7m_is_handler_mode(env) && spsel) { + return env->v7m.other_ss_psp; + } else { + return env->v7m.other_ss_msp; + } + } + default: + break; + } + } + switch (reg) { case 8: /* MSP */ return (env->v7m.control[env->v7m.secure] & R_V7M_CONTROL_SPSEL_MASK) ? @@ -XXX,XX +XXX,XX @@ void HELPER(v7m_msr)(CPUARMState *env, uint32_t maskreg, uint32_t val) return; } + if (arm_feature(env, ARM_FEATURE_M_SECURITY)) { + switch (reg) { + case 0x88: /* MSP_NS */ + if (!env->v7m.secure) { + return; + } + env->v7m.other_ss_msp = val; + return; + case 0x89: /* PSP_NS */ + if (!env->v7m.secure) { + return; + } + env->v7m.other_ss_psp = val; + return; + case 0x90: /* PRIMASK_NS */ + if (!env->v7m.secure) { + return; + } + env->v7m.primask[M_REG_NS] = val & 1; + return; + case 0x91: /* BASEPRI_NS */ + if (!env->v7m.secure) { + return; + } + env->v7m.basepri[M_REG_NS] = val & 0xff; + return; + case 0x93: /* FAULTMASK_NS */ + if (!env->v7m.secure) { + return; + } + env->v7m.faultmask[M_REG_NS] = val & 1; + return; + case 0x98: /* SP_NS */ + { + /* This gives the non-secure SP selected based on whether we're + * currently in handler mode or not, using the NS CONTROL.SPSEL. + */ + bool spsel = env->v7m.control[M_REG_NS] & R_V7M_CONTROL_SPSEL_MASK; + + if (!env->v7m.secure) { + return; + } + if (!arm_v7m_is_handler_mode(env) && spsel) { + env->v7m.other_ss_psp = val; + } else { + env->v7m.other_ss_msp = val; + } + return; + } + default: + break; + } + } + switch (reg) { case 0 ... 7: /* xPSR sub-fields */ /* only APSR is actually writable */ -- 2.7.4 For the v8M security extension, some exceptions must be banked between security states. Add the new vecinfo array which holds the state for the banked exceptions and migrate it if the CPU the NVIC is attached to implements the security extension. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- include/hw/intc/armv7m_nvic.h | 14 ++++++++++++ hw/intc/armv7m_nvic.c | 53 ++++++++++++++++++++++++++++++++++++++++++- 2 files changed, 66 insertions(+), 1 deletion(-) diff --git a/include/hw/intc/armv7m_nvic.h b/include/hw/intc/armv7m_nvic.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/intc/armv7m_nvic.h +++ b/include/hw/intc/armv7m_nvic.h @@ -XXX,XX +XXX,XX @@ /* Highest permitted number of exceptions (architectural limit) */ #define NVIC_MAX_VECTORS 512 +/* Number of internal exceptions */ +#define NVIC_INTERNAL_VECTORS 16 typedef struct VecInfo { /* Exception priorities can range from -3 to 255; only the unmodifiable @@ -XXX,XX +XXX,XX @@ typedef struct NVICState { ARMCPU *cpu; VecInfo vectors[NVIC_MAX_VECTORS]; + /* If the v8M security extension is implemented, some of the internal + * exceptions are banked between security states (ie there exists both + * a Secure and a NonSecure version of the exception and its state): + * HardFault, MemManage, UsageFault, SVCall, PendSV, SysTick (R_PJHV) + * The rest (including all the external exceptions) are not banked, though + * they may be configurable to target either Secure or NonSecure state. + * We store the secure exception state in sec_vectors[] for the banked + * exceptions, and otherwise use only vectors[] (including for exceptions + * like SecureFault that unconditionally target Secure state). + * Entries in sec_vectors[] for non-banked exception numbers are unused. + */ + VecInfo sec_vectors[NVIC_INTERNAL_VECTORS]; uint32_t prigroup; /* vectpending and exception_prio are both cached state that can diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ * For historical reasons QEMU tends to use "interrupt" and * "exception" more or less interchangeably. */ -#define NVIC_FIRST_IRQ 16 +#define NVIC_FIRST_IRQ NVIC_INTERNAL_VECTORS #define NVIC_MAX_IRQ (NVIC_MAX_VECTORS - NVIC_FIRST_IRQ) /* Effective running priority of the CPU when no exception is active @@ -XXX,XX +XXX,XX @@ static const VMStateDescription vmstate_VecInfo = { } }; +static bool nvic_security_needed(void *opaque) +{ + NVICState *s = opaque; + + return arm_feature(&s->cpu->env, ARM_FEATURE_M_SECURITY); +} + +static int nvic_security_post_load(void *opaque, int version_id) +{ + NVICState *s = opaque; + int i; + + /* Check for out of range priority settings */ + if (s->sec_vectors[ARMV7M_EXCP_HARD].prio != -1) { + return 1; + } + for (i = ARMV7M_EXCP_MEM; i < ARRAY_SIZE(s->sec_vectors); i++) { + if (s->sec_vectors[i].prio & ~0xff) { + return 1; + } + } + return 0; +} + +static const VMStateDescription vmstate_nvic_security = { + .name = "nvic/m-security", + .version_id = 1, + .minimum_version_id = 1, + .needed = nvic_security_needed, + .post_load = &nvic_security_post_load, + .fields = (VMStateField[]) { + VMSTATE_STRUCT_ARRAY(sec_vectors, NVICState, NVIC_INTERNAL_VECTORS, 1, + vmstate_VecInfo, VecInfo), + VMSTATE_END_OF_LIST() + } +}; + static const VMStateDescription vmstate_nvic = { .name = "armv7m_nvic", .version_id = 4, @@ -XXX,XX +XXX,XX @@ static const VMStateDescription vmstate_nvic = { vmstate_VecInfo, VecInfo), VMSTATE_UINT32(prigroup, NVICState), VMSTATE_END_OF_LIST() + }, + .subsections = (const VMStateDescription*[]) { + &vmstate_nvic_security, + NULL } }; @@ -XXX,XX +XXX,XX @@ static void armv7m_nvic_reset(DeviceState *dev) s->vectors[ARMV7M_EXCP_NMI].prio = -2; s->vectors[ARMV7M_EXCP_HARD].prio = -1; + if (arm_feature(&s->cpu->env, ARM_FEATURE_M_SECURITY)) { + s->sec_vectors[ARMV7M_EXCP_HARD].enabled = 1; + s->sec_vectors[ARMV7M_EXCP_SVC].enabled = 1; + s->sec_vectors[ARMV7M_EXCP_PENDSV].enabled = 1; + s->sec_vectors[ARMV7M_EXCP_SYSTICK].enabled = 1; + + /* AIRCR.BFHFNMINS resets to 0 so Secure HF is priority -1 (R_CMTC) */ + s->sec_vectors[ARMV7M_EXCP_HARD].prio = -1; + } + /* Strictly speaking the reset handler should be enabled. * However, we don't simulate soft resets through the NVIC, * and the reset vector should never be pended. -- 2.7.4 With banked exceptions, just the exception number in s->vectpending is no longer sufficient to uniquely identify the pending exception. Add a vectpending_is_s_banked bool which is true if the exception is using the sec_vectors[] array. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 1505240046-11454-4-git-send-email-peter.maydell@linaro.org --- include/hw/intc/armv7m_nvic.h | 11 +++++++++-- hw/intc/armv7m_nvic.c | 1 + 2 files changed, 10 insertions(+), 2 deletions(-) diff --git a/include/hw/intc/armv7m_nvic.h b/include/hw/intc/armv7m_nvic.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/intc/armv7m_nvic.h +++ b/include/hw/intc/armv7m_nvic.h @@ -XXX,XX +XXX,XX @@ typedef struct NVICState { VecInfo sec_vectors[NVIC_INTERNAL_VECTORS]; uint32_t prigroup; - /* vectpending and exception_prio are both cached state that can - * be recalculated from the vectors[] array and the prigroup field. + /* The following fields are all cached state that can be recalculated + * from the vectors[] and sec_vectors[] arrays and the prigroup field: + * - vectpending + * - vectpending_is_secure + * - exception_prio */ unsigned int vectpending; /* highest prio pending enabled exception */ + /* true if vectpending is a banked secure exception, ie it is in + * sec_vectors[] rather than vectors[] + */ + bool vectpending_is_s_banked; int exception_prio; /* group prio of the highest prio active exception */ MemoryRegion sysregmem; diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ static void armv7m_nvic_reset(DeviceState *dev) s->exception_prio = NVIC_NOEXC_PRIO; s->vectpending = 0; + s->vectpending_is_s_banked = false; } static void nvic_systick_trigger(void *opaque, int n, int level) -- 2.7.4 Instead of looking up the pending priority in nvic_pending_prio(), cache it in a new state struct field. The calculation of the pending priority given the interrupt number is more complicated in v8M with the security extension, so the caching will be worthwhile. This changes nvic_pending_prio() from returning a full (group + subpriority) priority value to returning a group priority. This doesn't require changes to its callsites because we use it only in comparisons of the form execution_prio > nvic_pending_prio() and execution priority is always a group priority, so a test (exec prio > full prio) is true if and only if (execprio > group_prio). (Architecturally the expected comparison is with the group priority for this sort of "would we preempt" test; we were only doing a test with a full priority as an optimisation to avoid the mask, which is possible precisely because the two comparisons always give the same answer.) Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-5-git-send-email-peter.maydell@linaro.org --- include/hw/intc/armv7m_nvic.h | 2 ++ hw/intc/armv7m_nvic.c | 23 +++++++++++++---------- hw/intc/trace-events | 2 +- 3 files changed, 16 insertions(+), 11 deletions(-) diff --git a/include/hw/intc/armv7m_nvic.h b/include/hw/intc/armv7m_nvic.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/intc/armv7m_nvic.h +++ b/include/hw/intc/armv7m_nvic.h @@ -XXX,XX +XXX,XX @@ typedef struct NVICState { * - vectpending * - vectpending_is_secure * - exception_prio + * - vectpending_prio */ unsigned int vectpending; /* highest prio pending enabled exception */ /* true if vectpending is a banked secure exception, ie it is in @@ -XXX,XX +XXX,XX @@ typedef struct NVICState { */ bool vectpending_is_s_banked; int exception_prio; /* group prio of the highest prio active exception */ + int vectpending_prio; /* group prio of the exeception in vectpending */ MemoryRegion sysregmem; MemoryRegion sysreg_ns_mem; diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ static const uint8_t nvic_id[] = { static int nvic_pending_prio(NVICState *s) { - /* return the priority of the current pending interrupt, + /* return the group priority of the current pending interrupt, * or NVIC_NOEXC_PRIO if no interrupt is pending */ - return s->vectpending ? s->vectors[s->vectpending].prio : NVIC_NOEXC_PRIO; + return s->vectpending_prio; } /* Return the value of the ISCR RETTOBASE bit: @@ -XXX,XX +XXX,XX @@ static void nvic_recompute_state(NVICState *s) active_prio &= nvic_gprio_mask(s); } + if (pend_prio > 0) { + pend_prio &= nvic_gprio_mask(s); + } + s->vectpending = pend_irq; + s->vectpending_prio = pend_prio; s->exception_prio = active_prio; - trace_nvic_recompute_state(s->vectpending, s->exception_prio); + trace_nvic_recompute_state(s->vectpending, + s->vectpending_prio, + s->exception_prio); } /* Return the current execution priority of the CPU @@ -XXX,XX +XXX,XX @@ void armv7m_nvic_acknowledge_irq(void *opaque) CPUARMState *env = &s->cpu->env; const int pending = s->vectpending; const int running = nvic_exec_prio(s); - int pendgroupprio; VecInfo *vec; assert(pending > ARMV7M_EXCP_RESET && pending < s->num_irq); @@ -XXX,XX +XXX,XX @@ void armv7m_nvic_acknowledge_irq(void *opaque) assert(vec->enabled); assert(vec->pending); - pendgroupprio = vec->prio; - if (pendgroupprio > 0) { - pendgroupprio &= nvic_gprio_mask(s); - } - assert(pendgroupprio < running); + assert(s->vectpending_prio < running); - trace_nvic_acknowledge_irq(pending, vec->prio); + trace_nvic_acknowledge_irq(pending, s->vectpending_prio); vec->active = 1; vec->pending = 0; @@ -XXX,XX +XXX,XX @@ static void armv7m_nvic_reset(DeviceState *dev) s->exception_prio = NVIC_NOEXC_PRIO; s->vectpending = 0; s->vectpending_is_s_banked = false; + s->vectpending_prio = NVIC_NOEXC_PRIO; } static void nvic_systick_trigger(void *opaque, int n, int level) diff --git a/hw/intc/trace-events b/hw/intc/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/trace-events +++ b/hw/intc/trace-events @@ -XXX,XX +XXX,XX @@ gicv3_redist_set_irq(uint32_t cpu, int irq, int level) "GICv3 redistributor 0x%x gicv3_redist_send_sgi(uint32_t cpu, int irq) "GICv3 redistributor 0x%x pending SGI %d" # hw/intc/armv7m_nvic.c -nvic_recompute_state(int vectpending, int exception_prio) "NVIC state recomputed: vectpending %d exception_prio %d" +nvic_recompute_state(int vectpending, int vectpending_prio, int exception_prio) "NVIC state recomputed: vectpending %d vectpending_prio %d exception_prio %d" nvic_set_prio(int irq, uint8_t prio) "NVIC set irq %d priority %d" nvic_irq_update(int vectpending, int pendprio, int exception_prio, int level) "NVIC vectpending %d pending prio %d exception_prio %d: setting irq line to %d" nvic_escalate_prio(int irq, int irqprio, int runprio) "NVIC escalating irq %d to HardFault: insufficient priority %d >= %d" -- 2.7.4 The Application Interrupt and Reset Control Register has some changes for v8M: * new bits SYSRESETREQS, BFHFNMINS and PRIS: these all have real state if the security extension is implemented and otherwise are constant * the PRIGROUP field is banked between security states * non-secure code can be blocked from using the SYSRESET bit to reset the system if SYSRESETREQS is set Implement the new state and the changes to register read and write. For the moment we ignore the effects of the secure PRIGROUP. We will implement the effects of PRIS and BFHFNMIS later. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-6-git-send-email-peter.maydell@linaro.org --- include/hw/intc/armv7m_nvic.h | 3 ++- target/arm/cpu.h | 12 +++++++++++ hw/intc/armv7m_nvic.c | 49 +++++++++++++++++++++++++++++++++---------- target/arm/cpu.c | 7 +++++++ 4 files changed, 59 insertions(+), 12 deletions(-) diff --git a/include/hw/intc/armv7m_nvic.h b/include/hw/intc/armv7m_nvic.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/intc/armv7m_nvic.h +++ b/include/hw/intc/armv7m_nvic.h @@ -XXX,XX +XXX,XX @@ typedef struct NVICState { * Entries in sec_vectors[] for non-banked exception numbers are unused. */ VecInfo sec_vectors[NVIC_INTERNAL_VECTORS]; - uint32_t prigroup; + /* The PRIGROUP field in AIRCR is banked */ + uint32_t prigroup[M_REG_NUM_BANKS]; /* The following fields are all cached state that can be recalculated * from the vectors[] and sec_vectors[] arrays and the prigroup field: 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 @@ typedef struct CPUARMState { int exception; uint32_t primask[M_REG_NUM_BANKS]; uint32_t faultmask[M_REG_NUM_BANKS]; + uint32_t aircr; /* only holds r/w state if security extn implemented */ uint32_t secure; /* Is CPU in Secure state? (not guest visible) */ } v7m; @@ -XXX,XX +XXX,XX @@ FIELD(V7M_CCR, STKALIGN, 9, 1) FIELD(V7M_CCR, DC, 16, 1) FIELD(V7M_CCR, IC, 17, 1) +/* V7M AIRCR bits */ +FIELD(V7M_AIRCR, VECTRESET, 0, 1) +FIELD(V7M_AIRCR, VECTCLRACTIVE, 1, 1) +FIELD(V7M_AIRCR, SYSRESETREQ, 2, 1) +FIELD(V7M_AIRCR, SYSRESETREQS, 3, 1) +FIELD(V7M_AIRCR, PRIGROUP, 8, 3) +FIELD(V7M_AIRCR, BFHFNMINS, 13, 1) +FIELD(V7M_AIRCR, PRIS, 14, 1) +FIELD(V7M_AIRCR, ENDIANNESS, 15, 1) +FIELD(V7M_AIRCR, VECTKEY, 16, 16) + /* V7M CFSR bits for MMFSR */ FIELD(V7M_CFSR, IACCVIOL, 0, 1) FIELD(V7M_CFSR, DACCVIOL, 1, 1) diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ static bool nvic_isrpending(NVICState *s) */ static inline uint32_t nvic_gprio_mask(NVICState *s) { - return ~0U << (s->prigroup + 1); + return ~0U << (s->prigroup[M_REG_NS] + 1); } /* Recompute vectpending and exception_prio */ @@ -XXX,XX +XXX,XX @@ static uint32_t nvic_readl(NVICState *s, uint32_t offset, MemTxAttrs attrs) return val; case 0xd08: /* Vector Table Offset. */ return cpu->env.v7m.vecbase[attrs.secure]; - case 0xd0c: /* Application Interrupt/Reset Control. */ - return 0xfa050000 | (s->prigroup << 8); + case 0xd0c: /* Application Interrupt/Reset Control (AIRCR) */ + val = 0xfa050000 | (s->prigroup[attrs.secure] << 8); + if (attrs.secure) { + /* s->aircr stores PRIS, BFHFNMINS, SYSRESETREQS */ + val |= cpu->env.v7m.aircr; + } else { + if (arm_feature(&cpu->env, ARM_FEATURE_V8)) { + /* BFHFNMINS is R/O from NS; other bits are RAZ/WI. If + * security isn't supported then BFHFNMINS is RAO (and + * the bit in env.v7m.aircr is always set). + */ + val |= cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK; + } + } + return val; case 0xd10: /* System Control. */ /* TODO: Implement SLEEPONEXIT. */ return 0; @@ -XXX,XX +XXX,XX @@ static void nvic_writel(NVICState *s, uint32_t offset, uint32_t value, case 0xd08: /* Vector Table Offset. */ cpu->env.v7m.vecbase[attrs.secure] = value & 0xffffff80; break; - case 0xd0c: /* Application Interrupt/Reset Control. */ - if ((value >> 16) == 0x05fa) { - if (value & 4) { - qemu_irq_pulse(s->sysresetreq); + case 0xd0c: /* Application Interrupt/Reset Control (AIRCR) */ + if ((value >> R_V7M_AIRCR_VECTKEY_SHIFT) == 0x05fa) { + if (value & R_V7M_AIRCR_SYSRESETREQ_MASK) { + if (attrs.secure || + !(cpu->env.v7m.aircr & R_V7M_AIRCR_SYSRESETREQS_MASK)) { + qemu_irq_pulse(s->sysresetreq); + } } - if (value & 2) { + if (value & R_V7M_AIRCR_VECTCLRACTIVE_MASK) { qemu_log_mask(LOG_GUEST_ERROR, "Setting VECTCLRACTIVE when not in DEBUG mode " "is UNPREDICTABLE\n"); } - if (value & 1) { + if (value & R_V7M_AIRCR_VECTRESET_MASK) { + /* NB: this bit is RES0 in v8M */ qemu_log_mask(LOG_GUEST_ERROR, "Setting VECTRESET when not in DEBUG mode " "is UNPREDICTABLE\n"); } - s->prigroup = extract32(value, 8, 3); + s->prigroup[attrs.secure] = extract32(value, + R_V7M_AIRCR_PRIGROUP_SHIFT, + R_V7M_AIRCR_PRIGROUP_LENGTH); + if (attrs.secure) { + /* These bits are only writable by secure */ + cpu->env.v7m.aircr = value & + (R_V7M_AIRCR_SYSRESETREQS_MASK | + R_V7M_AIRCR_BFHFNMINS_MASK | + R_V7M_AIRCR_PRIS_MASK); + } nvic_irq_update(s); } break; @@ -XXX,XX +XXX,XX @@ static const VMStateDescription vmstate_nvic_security = { .fields = (VMStateField[]) { VMSTATE_STRUCT_ARRAY(sec_vectors, NVICState, NVIC_INTERNAL_VECTORS, 1, vmstate_VecInfo, VecInfo), + VMSTATE_UINT32(prigroup[M_REG_S], NVICState), VMSTATE_END_OF_LIST() } }; @@ -XXX,XX +XXX,XX @@ static const VMStateDescription vmstate_nvic = { .fields = (VMStateField[]) { VMSTATE_STRUCT_ARRAY(vectors, NVICState, NVIC_MAX_VECTORS, 1, vmstate_VecInfo, VecInfo), - VMSTATE_UINT32(prigroup, NVICState), + VMSTATE_UINT32(prigroup[M_REG_NS], NVICState), VMSTATE_END_OF_LIST() }, .subsections = (const VMStateDescription*[]) { diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ static void arm_cpu_reset(CPUState *s) if (arm_feature(env, ARM_FEATURE_M_SECURITY)) { env->v7m.secure = true; + } else { + /* This bit resets to 0 if security is supported, but 1 if + * it is not. The bit is not present in v7M, but we set it + * here so we can avoid having to make checks on it conditional + * on ARM_FEATURE_V8 (we don't let the guest see the bit). + */ + env->v7m.aircr = R_V7M_AIRCR_BFHFNMINS_MASK; } /* In v7M the reset value of this bit is IMPDEF, but ARM recommends -- 2.7.4 Update the code in nvic_rettobase() so that it checks the sec_vectors[] array as well as the vectors[] array if needed. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-7-git-send-email-peter.maydell@linaro.org --- hw/intc/armv7m_nvic.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ static int nvic_pending_prio(NVICState *s) static bool nvic_rettobase(NVICState *s) { int irq, nhand = 0; + bool check_sec = arm_feature(&s->cpu->env, ARM_FEATURE_M_SECURITY); for (irq = ARMV7M_EXCP_RESET; irq < s->num_irq; irq++) { - if (s->vectors[irq].active) { + if (s->vectors[irq].active || + (check_sec && irq < NVIC_INTERNAL_VECTORS && + s->sec_vectors[irq].active)) { nhand++; if (nhand == 2) { return 0; -- 2.7.4 For v8M, the NVIC has a new set of registers per interrupt, NVIC_ITNS<n>. These determine whether the interrupt targets Secure or Non-secure state. Implement the register read/write code for these, and make them cause NVIC_IABR, NVIC_ICER, NVIC_ISER, NVIC_ICPR, NVIC_IPR and NVIC_ISPR to RAZ/WI for non-secure accesses to fields corresponding to interrupts which are configured to target secure state. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-8-git-send-email-peter.maydell@linaro.org --- include/hw/intc/armv7m_nvic.h | 3 ++ hw/intc/armv7m_nvic.c | 74 +++++++++++++++++++++++++++++++++++++++---- 2 files changed, 70 insertions(+), 7 deletions(-) diff --git a/include/hw/intc/armv7m_nvic.h b/include/hw/intc/armv7m_nvic.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/intc/armv7m_nvic.h +++ b/include/hw/intc/armv7m_nvic.h @@ -XXX,XX +XXX,XX @@ typedef struct NVICState { /* The PRIGROUP field in AIRCR is banked */ uint32_t prigroup[M_REG_NUM_BANKS]; + /* v8M NVIC_ITNS state (stored as a bool per bit) */ + bool itns[NVIC_MAX_VECTORS]; + /* The following fields are all cached state that can be recalculated * from the vectors[] and sec_vectors[] arrays and the prigroup field: * - vectpending diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ static uint32_t nvic_readl(NVICState *s, uint32_t offset, MemTxAttrs attrs) switch (offset) { case 4: /* Interrupt Control Type. */ return ((s->num_irq - NVIC_FIRST_IRQ) / 32) - 1; + case 0x380 ... 0x3bf: /* NVIC_ITNS<n> */ + { + int startvec = 32 * (offset - 0x380) + NVIC_FIRST_IRQ; + int i; + + if (!arm_feature(&cpu->env, ARM_FEATURE_V8)) { + goto bad_offset; + } + if (!attrs.secure) { + return 0; + } + val = 0; + for (i = 0; i < 32 && startvec + i < s->num_irq; i++) { + if (s->itns[startvec + i]) { + val |= (1 << i); + } + } + return val; + } case 0xd00: /* CPUID Base. */ return cpu->midr; case 0xd04: /* Interrupt Control State. */ @@ -XXX,XX +XXX,XX @@ static void nvic_writel(NVICState *s, uint32_t offset, uint32_t value, ARMCPU *cpu = s->cpu; switch (offset) { + case 0x380 ... 0x3bf: /* NVIC_ITNS<n> */ + { + int startvec = 32 * (offset - 0x380) + NVIC_FIRST_IRQ; + int i; + + if (!arm_feature(&cpu->env, ARM_FEATURE_V8)) { + goto bad_offset; + } + if (!attrs.secure) { + break; + } + for (i = 0; i < 32 && startvec + i < s->num_irq; i++) { + s->itns[startvec + i] = (value >> i) & 1; + } + nvic_irq_update(s); + break; + } case 0xd04: /* Interrupt Control State. */ if (value & (1 << 31)) { armv7m_nvic_set_pending(s, ARMV7M_EXCP_NMI); @@ -XXX,XX +XXX,XX @@ static MemTxResult nvic_sysreg_read(void *opaque, hwaddr addr, startvec = offset - 0x180 + NVIC_FIRST_IRQ; /* vector # */ for (i = 0, end = size * 8; i < end && startvec + i < s->num_irq; i++) { - if (s->vectors[startvec + i].enabled) { + if (s->vectors[startvec + i].enabled && + (attrs.secure || s->itns[startvec + i])) { val |= (1 << i); } } @@ -XXX,XX +XXX,XX @@ static MemTxResult nvic_sysreg_read(void *opaque, hwaddr addr, val = 0; startvec = offset - 0x280 + NVIC_FIRST_IRQ; /* vector # */ for (i = 0, end = size * 8; i < end && startvec + i < s->num_irq; i++) { - if (s->vectors[startvec + i].pending) { + if (s->vectors[startvec + i].pending && + (attrs.secure || s->itns[startvec + i])) { val |= (1 << i); } } @@ -XXX,XX +XXX,XX @@ static MemTxResult nvic_sysreg_read(void *opaque, hwaddr addr, startvec = offset - 0x300 + NVIC_FIRST_IRQ; /* vector # */ for (i = 0, end = size * 8; i < end && startvec + i < s->num_irq; i++) { - if (s->vectors[startvec + i].active) { + if (s->vectors[startvec + i].active && + (attrs.secure || s->itns[startvec + i])) { val |= (1 << i); } } @@ -XXX,XX +XXX,XX @@ static MemTxResult nvic_sysreg_read(void *opaque, hwaddr addr, startvec = offset - 0x400 + NVIC_FIRST_IRQ; /* vector # */ for (i = 0; i < size && startvec + i < s->num_irq; i++) { - val |= s->vectors[startvec + i].prio << (8 * i); + if (attrs.secure || s->itns[startvec + i]) { + val |= s->vectors[startvec + i].prio << (8 * i); + } } break; case 0xd18 ... 0xd23: /* System Handler Priority. */ @@ -XXX,XX +XXX,XX @@ static MemTxResult nvic_sysreg_write(void *opaque, hwaddr addr, startvec = 8 * (offset - 0x180) + NVIC_FIRST_IRQ; for (i = 0, end = size * 8; i < end && startvec + i < s->num_irq; i++) { - if (value & (1 << i)) { + if (value & (1 << i) && + (attrs.secure || s->itns[startvec + i])) { s->vectors[startvec + i].enabled = setval; } } @@ -XXX,XX +XXX,XX @@ static MemTxResult nvic_sysreg_write(void *opaque, hwaddr addr, startvec = 8 * (offset - 0x280) + NVIC_FIRST_IRQ; /* vector # */ for (i = 0, end = size * 8; i < end && startvec + i < s->num_irq; i++) { - if (value & (1 << i)) { + if (value & (1 << i) && + (attrs.secure || s->itns[startvec + i])) { s->vectors[startvec + i].pending = setval; } } @@ -XXX,XX +XXX,XX @@ static MemTxResult nvic_sysreg_write(void *opaque, hwaddr addr, startvec = 8 * (offset - 0x400) + NVIC_FIRST_IRQ; /* vector # */ for (i = 0; i < size && startvec + i < s->num_irq; i++) { - set_prio(s, startvec + i, (value >> (i * 8)) & 0xff); + if (attrs.secure || s->itns[startvec + i]) { + set_prio(s, startvec + i, (value >> (i * 8)) & 0xff); + } } nvic_irq_update(s); return MEMTX_OK; @@ -XXX,XX +XXX,XX @@ static const VMStateDescription vmstate_nvic_security = { VMSTATE_STRUCT_ARRAY(sec_vectors, NVICState, NVIC_INTERNAL_VECTORS, 1, vmstate_VecInfo, VecInfo), VMSTATE_UINT32(prigroup[M_REG_S], NVICState), + VMSTATE_BOOL_ARRAY(itns, NVICState, NVIC_MAX_VECTORS), VMSTATE_END_OF_LIST() } }; @@ -XXX,XX +XXX,XX @@ static void armv7m_nvic_reset(DeviceState *dev) s->vectpending = 0; s->vectpending_is_s_banked = false; s->vectpending_prio = NVIC_NOEXC_PRIO; + + if (arm_feature(&s->cpu->env, ARM_FEATURE_M_SECURITY)) { + memset(s->itns, 0, sizeof(s->itns)); + } else { + /* This state is constant and not guest accessible in a non-security + * NVIC; we set the bits to true to avoid having to do a feature + * bit check in the NVIC enable/pend/etc register accessors. + */ + int i; + + for (i = NVIC_FIRST_IRQ; i < ARRAY_SIZE(s->itns); i++) { + s->itns[i] = true; + } + } } static void nvic_systick_trigger(void *opaque, int n, int level) -- 2.7.4 Update the nvic_recompute_state() code to handle the security extension and its associated banked registers. Code that uses the resulting cached state (ie the irq acknowledge and complete code) will be updated in a later commit. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-9-git-send-email-peter.maydell@linaro.org --- hw/intc/armv7m_nvic.c | 151 ++++++++++++++++++++++++++++++++++++++++++++++++-- hw/intc/trace-events | 1 + 2 files changed, 147 insertions(+), 5 deletions(-) diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ * (higher than the highest possible priority value) */ #define NVIC_NOEXC_PRIO 0x100 +/* Maximum priority of non-secure exceptions when AIRCR.PRIS is set */ +#define NVIC_NS_PRIO_LIMIT 0x80 static const uint8_t nvic_id[] = { 0x00, 0xb0, 0x1b, 0x00, 0x0d, 0xe0, 0x05, 0xb1 @@ -XXX,XX +XXX,XX @@ static bool nvic_isrpending(NVICState *s) return false; } +static bool exc_is_banked(int exc) +{ + /* Return true if this is one of the limited set of exceptions which + * are banked (and thus have state in sec_vectors[]) + */ + return exc == ARMV7M_EXCP_HARD || + exc == ARMV7M_EXCP_MEM || + exc == ARMV7M_EXCP_USAGE || + exc == ARMV7M_EXCP_SVC || + exc == ARMV7M_EXCP_PENDSV || + exc == ARMV7M_EXCP_SYSTICK; +} + /* Return a mask word which clears the subpriority bits from * a priority value for an M-profile exception, leaving only * the group priority. */ -static inline uint32_t nvic_gprio_mask(NVICState *s) +static inline uint32_t nvic_gprio_mask(NVICState *s, bool secure) +{ + return ~0U << (s->prigroup[secure] + 1); +} + +static bool exc_targets_secure(NVICState *s, int exc) +{ + /* Return true if this non-banked exception targets Secure state. */ + if (!arm_feature(&s->cpu->env, ARM_FEATURE_M_SECURITY)) { + return false; + } + + if (exc >= NVIC_FIRST_IRQ) { + return !s->itns[exc]; + } + + /* Function shouldn't be called for banked exceptions. */ + assert(!exc_is_banked(exc)); + + switch (exc) { + case ARMV7M_EXCP_NMI: + case ARMV7M_EXCP_BUS: + return !(s->cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK); + case ARMV7M_EXCP_SECURE: + return true; + case ARMV7M_EXCP_DEBUG: + /* TODO: controlled by DEMCR.SDME, which we don't yet implement */ + return false; + default: + /* reset, and reserved (unused) low exception numbers. + * We'll get called by code that loops through all the exception + * numbers, but it doesn't matter what we return here as these + * non-existent exceptions will never be pended or active. + */ + return true; + } +} + +static int exc_group_prio(NVICState *s, int rawprio, bool targets_secure) +{ + /* Return the group priority for this exception, given its raw + * (group-and-subgroup) priority value and whether it is targeting + * secure state or not. + */ + if (rawprio < 0) { + return rawprio; + } + rawprio &= nvic_gprio_mask(s, targets_secure); + /* AIRCR.PRIS causes us to squash all NS priorities into the + * lower half of the total range + */ + if (!targets_secure && + (s->cpu->env.v7m.aircr & R_V7M_AIRCR_PRIS_MASK)) { + rawprio = (rawprio >> 1) + NVIC_NS_PRIO_LIMIT; + } + return rawprio; +} + +/* Recompute vectpending and exception_prio for a CPU which implements + * the Security extension + */ +static void nvic_recompute_state_secure(NVICState *s) { - return ~0U << (s->prigroup[M_REG_NS] + 1); + int i, bank; + int pend_prio = NVIC_NOEXC_PRIO; + int active_prio = NVIC_NOEXC_PRIO; + int pend_irq = 0; + bool pending_is_s_banked = false; + + /* R_CQRV: precedence is by: + * - lowest group priority; if both the same then + * - lowest subpriority; if both the same then + * - lowest exception number; if both the same (ie banked) then + * - secure exception takes precedence + * Compare pseudocode RawExecutionPriority. + * Annoyingly, now we have two prigroup values (for S and NS) + * we can't do the loop comparison on raw priority values. + */ + for (i = 1; i < s->num_irq; i++) { + for (bank = M_REG_S; bank >= M_REG_NS; bank--) { + VecInfo *vec; + int prio; + bool targets_secure; + + if (bank == M_REG_S) { + if (!exc_is_banked(i)) { + continue; + } + vec = &s->sec_vectors[i]; + targets_secure = true; + } else { + vec = &s->vectors[i]; + targets_secure = !exc_is_banked(i) && exc_targets_secure(s, i); + } + + prio = exc_group_prio(s, vec->prio, targets_secure); + if (vec->enabled && vec->pending && prio < pend_prio) { + pend_prio = prio; + pend_irq = i; + pending_is_s_banked = (bank == M_REG_S); + } + if (vec->active && prio < active_prio) { + active_prio = prio; + } + } + } + + s->vectpending_is_s_banked = pending_is_s_banked; + s->vectpending = pend_irq; + s->vectpending_prio = pend_prio; + s->exception_prio = active_prio; + + trace_nvic_recompute_state_secure(s->vectpending, + s->vectpending_is_s_banked, + s->vectpending_prio, + s->exception_prio); } /* Recompute vectpending and exception_prio */ @@ -XXX,XX +XXX,XX @@ static void nvic_recompute_state(NVICState *s) int active_prio = NVIC_NOEXC_PRIO; int pend_irq = 0; + /* In theory we could write one function that handled both + * the "security extension present" and "not present"; however + * the security related changes significantly complicate the + * recomputation just by themselves and mixing both cases together + * would be even worse, so we retain a separate non-secure-only + * version for CPUs which don't implement the security extension. + */ + if (arm_feature(&s->cpu->env, ARM_FEATURE_M_SECURITY)) { + nvic_recompute_state_secure(s); + return; + } + for (i = 1; i < s->num_irq; i++) { VecInfo *vec = &s->vectors[i]; @@ -XXX,XX +XXX,XX @@ static void nvic_recompute_state(NVICState *s) } if (active_prio > 0) { - active_prio &= nvic_gprio_mask(s); + active_prio &= nvic_gprio_mask(s, false); } if (pend_prio > 0) { - pend_prio &= nvic_gprio_mask(s); + pend_prio &= nvic_gprio_mask(s, false); } s->vectpending = pend_irq; @@ -XXX,XX +XXX,XX @@ static inline int nvic_exec_prio(NVICState *s) } else if (env->v7m.primask[env->v7m.secure]) { running = 0; } else if (env->v7m.basepri[env->v7m.secure] > 0) { - running = env->v7m.basepri[env->v7m.secure] & nvic_gprio_mask(s); + running = env->v7m.basepri[env->v7m.secure] & + nvic_gprio_mask(s, env->v7m.secure); } else { running = NVIC_NOEXC_PRIO; /* lower than any possible priority */ } diff --git a/hw/intc/trace-events b/hw/intc/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/trace-events +++ b/hw/intc/trace-events @@ -XXX,XX +XXX,XX @@ gicv3_redist_send_sgi(uint32_t cpu, int irq) "GICv3 redistributor 0x%x pending S # hw/intc/armv7m_nvic.c nvic_recompute_state(int vectpending, int vectpending_prio, int exception_prio) "NVIC state recomputed: vectpending %d vectpending_prio %d exception_prio %d" +nvic_recompute_state_secure(int vectpending, bool vectpending_is_s_banked, int vectpending_prio, int exception_prio) "NVIC state recomputed: vectpending %d is_s_banked %d vectpending_prio %d exception_prio %d" nvic_set_prio(int irq, uint8_t prio) "NVIC set irq %d priority %d" nvic_irq_update(int vectpending, int pendprio, int exception_prio, int level) "NVIC vectpending %d pending prio %d exception_prio %d: setting irq line to %d" nvic_escalate_prio(int irq, int irqprio, int runprio) "NVIC escalating irq %d to HardFault: insufficient priority %d >= %d" -- 2.7.4 Make the armv7m_nvic_set_pending() and armv7m_nvic_clear_pending() functions take a bool indicating whether to pend the secure or non-secure version of a banked interrupt, and update the callsites accordingly. In most callsites we can simply pass the correct security state in; in a couple of cases we use TODO comments to indicate that we will return the code in a subsequent commit. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-10-git-send-email-peter.maydell@linaro.org --- target/arm/cpu.h | 14 ++++++++++- hw/intc/armv7m_nvic.c | 64 ++++++++++++++++++++++++++++++++++++++------------- target/arm/helper.c | 24 +++++++++++-------- hw/intc/trace-events | 4 ++-- 4 files changed, 77 insertions(+), 29 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 @@ static inline bool armv7m_nvic_can_take_pending_exception(void *opaque) return true; } #endif -void armv7m_nvic_set_pending(void *opaque, int irq); +/** + * armv7m_nvic_set_pending: mark the specified exception as pending + * @opaque: the NVIC + * @irq: the exception number to mark pending + * @secure: false for non-banked exceptions or for the nonsecure + * version of a banked exception, true for the secure version of a banked + * exception. + * + * Marks the specified exception as pending. Note that we will assert() + * if @secure is true and @irq does not specify one of the fixed set + * of architecturally banked exceptions. + */ +void armv7m_nvic_set_pending(void *opaque, int irq, bool secure); void armv7m_nvic_acknowledge_irq(void *opaque); /** * armv7m_nvic_complete_irq: complete specified interrupt or exception diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ static void nvic_irq_update(NVICState *s) qemu_set_irq(s->excpout, lvl); } -static void armv7m_nvic_clear_pending(void *opaque, int irq) +/** + * armv7m_nvic_clear_pending: mark the specified exception as not pending + * @opaque: the NVIC + * @irq: the exception number to mark as not pending + * @secure: false for non-banked exceptions or for the nonsecure + * version of a banked exception, true for the secure version of a banked + * exception. + * + * Marks the specified exception as not pending. Note that we will assert() + * if @secure is true and @irq does not specify one of the fixed set + * of architecturally banked exceptions. + */ +static void armv7m_nvic_clear_pending(void *opaque, int irq, bool secure) { NVICState *s = (NVICState *)opaque; VecInfo *vec; assert(irq > ARMV7M_EXCP_RESET && irq < s->num_irq); - vec = &s->vectors[irq]; - trace_nvic_clear_pending(irq, vec->enabled, vec->prio); + if (secure) { + assert(exc_is_banked(irq)); + vec = &s->sec_vectors[irq]; + } else { + vec = &s->vectors[irq]; + } + trace_nvic_clear_pending(irq, secure, vec->enabled, vec->prio); if (vec->pending) { vec->pending = 0; nvic_irq_update(s); } } -void armv7m_nvic_set_pending(void *opaque, int irq) +void armv7m_nvic_set_pending(void *opaque, int irq, bool secure) { NVICState *s = (NVICState *)opaque; + bool banked = exc_is_banked(irq); VecInfo *vec; assert(irq > ARMV7M_EXCP_RESET && irq < s->num_irq); + assert(!secure || banked); - vec = &s->vectors[irq]; - trace_nvic_set_pending(irq, vec->enabled, vec->prio); + vec = (banked && secure) ? &s->sec_vectors[irq] : &s->vectors[irq]; + trace_nvic_set_pending(irq, secure, vec->enabled, vec->prio); if (irq >= ARMV7M_EXCP_HARD && irq < ARMV7M_EXCP_PENDSV) { /* If a synchronous exception is pending then it may be @@ -XXX,XX +XXX,XX @@ void armv7m_nvic_set_pending(void *opaque, int irq) "(current priority %d)\n", irq, running); } - /* We can do the escalation, so we take HardFault instead */ + /* We can do the escalation, so we take HardFault instead. + * If BFHFNMINS is set then we escalate to the banked HF for + * the target security state of the original exception; otherwise + * we take a Secure HardFault. + */ irq = ARMV7M_EXCP_HARD; - vec = &s->vectors[irq]; + if (arm_feature(&s->cpu->env, ARM_FEATURE_M_SECURITY) && + (secure || + !(s->cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK))) { + vec = &s->sec_vectors[irq]; + } else { + vec = &s->vectors[irq]; + } + /* HF may be banked but there is only one shared HFSR */ s->cpu->env.v7m.hfsr |= R_V7M_HFSR_FORCED_MASK; } } @@ -XXX,XX +XXX,XX @@ static void set_irq_level(void *opaque, int n, int level) if (level != vec->level) { vec->level = level; if (level) { - armv7m_nvic_set_pending(s, n); + armv7m_nvic_set_pending(s, n, false); } } } @@ -XXX,XX +XXX,XX @@ static void nvic_writel(NVICState *s, uint32_t offset, uint32_t value, } case 0xd04: /* Interrupt Control State. */ if (value & (1 << 31)) { - armv7m_nvic_set_pending(s, ARMV7M_EXCP_NMI); + armv7m_nvic_set_pending(s, ARMV7M_EXCP_NMI, false); } if (value & (1 << 28)) { - armv7m_nvic_set_pending(s, ARMV7M_EXCP_PENDSV); + armv7m_nvic_set_pending(s, ARMV7M_EXCP_PENDSV, attrs.secure); } else if (value & (1 << 27)) { - armv7m_nvic_clear_pending(s, ARMV7M_EXCP_PENDSV); + armv7m_nvic_clear_pending(s, ARMV7M_EXCP_PENDSV, attrs.secure); } if (value & (1 << 26)) { - armv7m_nvic_set_pending(s, ARMV7M_EXCP_SYSTICK); + armv7m_nvic_set_pending(s, ARMV7M_EXCP_SYSTICK, attrs.secure); } else if (value & (1 << 25)) { - armv7m_nvic_clear_pending(s, ARMV7M_EXCP_SYSTICK); + armv7m_nvic_clear_pending(s, ARMV7M_EXCP_SYSTICK, attrs.secure); } break; case 0xd08: /* Vector Table Offset. */ @@ -XXX,XX +XXX,XX @@ static void nvic_writel(NVICState *s, uint32_t offset, uint32_t value, { int excnum = (value & 0x1ff) + NVIC_FIRST_IRQ; if (excnum < s->num_irq) { - armv7m_nvic_set_pending(s, excnum); + armv7m_nvic_set_pending(s, excnum, false); } break; } @@ -XXX,XX +XXX,XX @@ static void nvic_systick_trigger(void *opaque, int n, int level) /* SysTick just asked us to pend its exception. * (This is different from an external interrupt line's * behaviour.) + * TODO: when we implement the banked systicks we must make + * this pend the correct banked exception. */ - armv7m_nvic_set_pending(s, ARMV7M_EXCP_SYSTICK); + armv7m_nvic_set_pending(s, ARMV7M_EXCP_SYSTICK, false); } } 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) * stack, directly take a usage fault on the current stack. */ env->v7m.cfsr[env->v7m.secure] |= R_V7M_CFSR_INVPC_MASK; - armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_USAGE); + armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_USAGE, env->v7m.secure); v7m_exception_taken(cpu, excret); qemu_log_mask(CPU_LOG_INT, "...taking UsageFault on existing " "stackframe: failed exception return integrity check\n"); @@ -XXX,XX +XXX,XX @@ static void do_v7m_exception_exit(ARMCPU *cpu) * exception return excret specified then this is a UsageFault. */ if (return_to_handler != arm_v7m_is_handler_mode(env)) { - /* Take an INVPC UsageFault by pushing the stack again. */ - armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_USAGE); + /* Take an INVPC UsageFault by pushing the stack again. + * TODO: the v8M version of this code should target the + * background state for this exception. + */ + armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_USAGE, false); env->v7m.cfsr[env->v7m.secure] |= R_V7M_CFSR_INVPC_MASK; v7m_push_stack(cpu); v7m_exception_taken(cpu, excret); @@ -XXX,XX +XXX,XX @@ void arm_v7m_cpu_do_interrupt(CPUState *cs) handle it. */ switch (cs->exception_index) { case EXCP_UDEF: - armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_USAGE); + armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_USAGE, env->v7m.secure); env->v7m.cfsr[env->v7m.secure] |= R_V7M_CFSR_UNDEFINSTR_MASK; break; case EXCP_NOCP: - armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_USAGE); + armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_USAGE, env->v7m.secure); env->v7m.cfsr[env->v7m.secure] |= R_V7M_CFSR_NOCP_MASK; break; case EXCP_INVSTATE: - armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_USAGE); + 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_SWI: /* The PC already points to the next instruction. */ - armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_SVC); + armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_SVC, env->v7m.secure); break; case EXCP_PREFETCH_ABORT: case EXCP_DATA_ABORT: @@ -XXX,XX +XXX,XX @@ void arm_v7m_cpu_do_interrupt(CPUState *cs) env->v7m.bfar); break; } - armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_BUS); + armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_BUS, false); break; default: /* All other FSR values are either MPU faults or "can't happen @@ -XXX,XX +XXX,XX @@ void arm_v7m_cpu_do_interrupt(CPUState *cs) env->v7m.mmfar[env->v7m.secure]); break; } - armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_MEM); + armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_MEM, + env->v7m.secure); break; } break; @@ -XXX,XX +XXX,XX @@ void arm_v7m_cpu_do_interrupt(CPUState *cs) return; } } - armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_DEBUG); + armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_DEBUG, false); break; case EXCP_IRQ: break; diff --git a/hw/intc/trace-events b/hw/intc/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/trace-events +++ b/hw/intc/trace-events @@ -XXX,XX +XXX,XX @@ nvic_set_prio(int irq, uint8_t prio) "NVIC set irq %d priority %d" nvic_irq_update(int vectpending, int pendprio, int exception_prio, int level) "NVIC vectpending %d pending prio %d exception_prio %d: setting irq line to %d" nvic_escalate_prio(int irq, int irqprio, int runprio) "NVIC escalating irq %d to HardFault: insufficient priority %d >= %d" nvic_escalate_disabled(int irq) "NVIC escalating irq %d to HardFault: disabled" -nvic_set_pending(int irq, int en, int prio) "NVIC set pending irq %d (enabled: %d priority %d)" -nvic_clear_pending(int irq, int en, int prio) "NVIC clear pending irq %d (enabled: %d priority %d)" +nvic_set_pending(int irq, bool secure, int en, int prio) "NVIC set pending irq %d secure-bank %d (enabled: %d priority %d)" +nvic_clear_pending(int irq, bool secure, int en, int prio) "NVIC clear pending irq %d secure-bank %d (enabled: %d priority %d)" nvic_set_pending_level(int irq) "NVIC set pending: irq %d higher prio than vectpending: setting irq line to 1" nvic_acknowledge_irq(int irq, int prio) "NVIC acknowledge IRQ: %d now active (prio %d)" nvic_complete_irq(int irq) "NVIC complete IRQ %d" -- 2.7.4 Make the set_prio() function take a bool indicating whether to pend the secure or non-secure version of a banked interrupt, and use this to implement the correct banking semantics for the SHPR registers. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-11-git-send-email-peter.maydell@linaro.org --- hw/intc/armv7m_nvic.c | 96 ++++++++++++++++++++++++++++++++++++++++++++++----- hw/intc/trace-events | 2 +- 2 files changed, 88 insertions(+), 10 deletions(-) diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ int armv7m_nvic_raw_execution_priority(void *opaque) return s->exception_prio; } -/* caller must call nvic_irq_update() after this */ -static void set_prio(NVICState *s, unsigned irq, uint8_t prio) +/* caller must call nvic_irq_update() after this. + * secure indicates the bank to use for banked exceptions (we assert if + * we are passed secure=true for a non-banked exception). + */ +static void set_prio(NVICState *s, unsigned irq, bool secure, uint8_t prio) { assert(irq > ARMV7M_EXCP_NMI); /* only use for configurable prios */ assert(irq < s->num_irq); - s->vectors[irq].prio = prio; + if (secure) { + assert(exc_is_banked(irq)); + s->sec_vectors[irq].prio = prio; + } else { + s->vectors[irq].prio = prio; + } + + trace_nvic_set_prio(irq, secure, prio); +} + +/* Return the current raw priority register value. + * secure indicates the bank to use for banked exceptions (we assert if + * we are passed secure=true for a non-banked exception). + */ +static int get_prio(NVICState *s, unsigned irq, bool secure) +{ + assert(irq > ARMV7M_EXCP_NMI); /* only use for configurable prios */ + assert(irq < s->num_irq); - trace_nvic_set_prio(irq, prio); + if (secure) { + assert(exc_is_banked(irq)); + return s->sec_vectors[irq].prio; + } else { + return s->vectors[irq].prio; + } } /* Recompute state and assert irq line accordingly. @@ -XXX,XX +XXX,XX @@ static bool nvic_user_access_ok(NVICState *s, hwaddr offset, MemTxAttrs attrs) } } +static int shpr_bank(NVICState *s, int exc, MemTxAttrs attrs) +{ + /* Behaviour for the SHPR register field for this exception: + * return M_REG_NS to use the nonsecure vector (including for + * non-banked exceptions), M_REG_S for the secure version of + * a banked exception, and -1 if this field should RAZ/WI. + */ + switch (exc) { + case ARMV7M_EXCP_MEM: + case ARMV7M_EXCP_USAGE: + case ARMV7M_EXCP_SVC: + case ARMV7M_EXCP_PENDSV: + case ARMV7M_EXCP_SYSTICK: + /* Banked exceptions */ + return attrs.secure; + case ARMV7M_EXCP_BUS: + /* Not banked, RAZ/WI from nonsecure if BFHFNMINS is zero */ + if (!attrs.secure && + !(s->cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK)) { + return -1; + } + return M_REG_NS; + case ARMV7M_EXCP_SECURE: + /* Not banked, RAZ/WI from nonsecure */ + if (!attrs.secure) { + return -1; + } + return M_REG_NS; + case ARMV7M_EXCP_DEBUG: + /* Not banked. TODO should RAZ/WI if DEMCR.SDME is set */ + return M_REG_NS; + case 8 ... 10: + case 13: + /* RES0 */ + return -1; + default: + /* Not reachable due to decode of SHPR register addresses */ + g_assert_not_reached(); + } +} + static MemTxResult nvic_sysreg_read(void *opaque, hwaddr addr, uint64_t *data, unsigned size, MemTxAttrs attrs) @@ -XXX,XX +XXX,XX @@ static MemTxResult nvic_sysreg_read(void *opaque, hwaddr addr, } } break; - case 0xd18 ... 0xd23: /* System Handler Priority. */ + case 0xd18 ... 0xd23: /* System Handler Priority (SHPR1, SHPR2, SHPR3) */ val = 0; for (i = 0; i < size; i++) { - val |= s->vectors[(offset - 0xd14) + i].prio << (i * 8); + unsigned hdlidx = (offset - 0xd14) + i; + int sbank = shpr_bank(s, hdlidx, attrs); + + if (sbank < 0) { + continue; + } + val = deposit32(val, i * 8, 8, get_prio(s, hdlidx, sbank)); } break; case 0xfe0 ... 0xfff: /* ID. */ @@ -XXX,XX +XXX,XX @@ static MemTxResult nvic_sysreg_write(void *opaque, hwaddr addr, for (i = 0; i < size && startvec + i < s->num_irq; i++) { if (attrs.secure || s->itns[startvec + i]) { - set_prio(s, startvec + i, (value >> (i * 8)) & 0xff); + set_prio(s, startvec + i, false, (value >> (i * 8)) & 0xff); } } nvic_irq_update(s); return MEMTX_OK; - case 0xd18 ... 0xd23: /* System Handler Priority. */ + case 0xd18 ... 0xd23: /* System Handler Priority (SHPR1, SHPR2, SHPR3) */ for (i = 0; i < size; i++) { unsigned hdlidx = (offset - 0xd14) + i; - set_prio(s, hdlidx, (value >> (i * 8)) & 0xff); + int newprio = extract32(value, i * 8, 8); + int sbank = shpr_bank(s, hdlidx, attrs); + + if (sbank < 0) { + continue; + } + set_prio(s, hdlidx, sbank, newprio); } nvic_irq_update(s); return MEMTX_OK; diff --git a/hw/intc/trace-events b/hw/intc/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/trace-events +++ b/hw/intc/trace-events @@ -XXX,XX +XXX,XX @@ gicv3_redist_send_sgi(uint32_t cpu, int irq) "GICv3 redistributor 0x%x pending S # hw/intc/armv7m_nvic.c nvic_recompute_state(int vectpending, int vectpending_prio, int exception_prio) "NVIC state recomputed: vectpending %d vectpending_prio %d exception_prio %d" nvic_recompute_state_secure(int vectpending, bool vectpending_is_s_banked, int vectpending_prio, int exception_prio) "NVIC state recomputed: vectpending %d is_s_banked %d vectpending_prio %d exception_prio %d" -nvic_set_prio(int irq, uint8_t prio) "NVIC set irq %d priority %d" +nvic_set_prio(int irq, bool secure, uint8_t prio) "NVIC set irq %d secure-bank %d priority %d" nvic_irq_update(int vectpending, int pendprio, int exception_prio, int level) "NVIC vectpending %d pending prio %d exception_prio %d: setting irq line to %d" nvic_escalate_prio(int irq, int irqprio, int runprio) "NVIC escalating irq %d to HardFault: insufficient priority %d >= %d" nvic_escalate_disabled(int irq) "NVIC escalating irq %d to HardFault: disabled" -- 2.7.4 In armv7m_nvic_set_pending() we have to compare the priority of an exception against the execution priority to decide whether it needs to be escalated to HardFault. In the specification this is a comparison against the exception's group priority; for v7M we implemented it as a comparison against the raw exception priority because the two comparisons will always give the same answer. For v8M the existence of AIRCR.PRIS and the possibility of different PRIGROUP values for secure and nonsecure exceptions means we need to explicitly calculate the vector's group priority for this check. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-12-git-send-email-peter.maydell@linaro.org --- hw/intc/armv7m_nvic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ void armv7m_nvic_set_pending(void *opaque, int irq, bool secure) int running = nvic_exec_prio(s); bool escalate = false; - if (vec->prio >= running) { + if (exc_group_prio(s, vec->prio, secure) >= running) { trace_nvic_escalate_prio(irq, vec->prio, running); escalate = true; } else if (!vec->enabled) { -- 2.7.4 When escalating to HardFault, we must go into Lockup if we can't take the synchronous HardFault because the current execution priority is already at or below the priority of HardFault. In v7M HF is always priority -1 so a simple < 0 comparison sufficed; in v8M the priority of HardFault can vary depending on whether it is a Secure or NonSecure HardFault, so we must check against the priority of the HardFault exception vector we're about to use. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-13-git-send-email-peter.maydell@linaro.org --- hw/intc/armv7m_nvic.c | 23 ++++++++++++----------- 1 file changed, 12 insertions(+), 11 deletions(-) diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ void armv7m_nvic_set_pending(void *opaque, int irq, bool secure) } if (escalate) { - if (running < 0) { - /* We want to escalate to HardFault but we can't take a - * synchronous HardFault at this point either. This is a - * Lockup condition due to a guest bug. We don't model - * Lockup, so report via cpu_abort() instead. - */ - cpu_abort(&s->cpu->parent_obj, - "Lockup: can't escalate %d to HardFault " - "(current priority %d)\n", irq, running); - } - /* We can do the escalation, so we take HardFault instead. + /* We need to escalate this exception to a synchronous HardFault. * If BFHFNMINS is set then we escalate to the banked HF for * the target security state of the original exception; otherwise * we take a Secure HardFault. @@ -XXX,XX +XXX,XX @@ void armv7m_nvic_set_pending(void *opaque, int irq, bool secure) } else { vec = &s->vectors[irq]; } + if (running <= vec->prio) { + /* We want to escalate to HardFault but we can't take the + * synchronous HardFault at this point either. This is a + * Lockup condition due to a guest bug. We don't model + * Lockup, so report via cpu_abort() instead. + */ + cpu_abort(&s->cpu->parent_obj, + "Lockup: can't escalate %d to HardFault " + "(current priority %d)\n", irq, running); + } + /* HF may be banked but there is only one shared HFSR */ s->cpu->env.v7m.hfsr |= R_V7M_HFSR_FORCED_MASK; } -- 2.7.4 In v7M, the fixed-priority exceptions are: Reset: -3 NMI: -2 HardFault: -1 In v8M, this changes because Secure HardFault may need to be prioritised above NMI: Reset: -4 Secure HardFault if AIRCR.BFHFNMINS == 1: -3 NMI: -2 Secure HardFault if AIRCR.BFHFNMINS == 0: -1 NonSecure HardFault: -1 Make these changes, including support for changing the priority of Secure HardFault as AIRCR.BFHFNMINS changes. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-14-git-send-email-peter.maydell@linaro.org --- hw/intc/armv7m_nvic.c | 22 +++++++++++++++++++--- 1 file changed, 19 insertions(+), 3 deletions(-) diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ static void nvic_writel(NVICState *s, uint32_t offset, uint32_t value, (R_V7M_AIRCR_SYSRESETREQS_MASK | R_V7M_AIRCR_BFHFNMINS_MASK | R_V7M_AIRCR_PRIS_MASK); + /* BFHFNMINS changes the priority of Secure HardFault */ + if (cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK) { + s->sec_vectors[ARMV7M_EXCP_HARD].prio = -3; + } else { + s->sec_vectors[ARMV7M_EXCP_HARD].prio = -1; + } } nvic_irq_update(s); } @@ -XXX,XX +XXX,XX @@ static int nvic_post_load(void *opaque, int version_id) { NVICState *s = opaque; unsigned i; + int resetprio; /* Check for out of range priority settings */ - if (s->vectors[ARMV7M_EXCP_RESET].prio != -3 || + resetprio = arm_feature(&s->cpu->env, ARM_FEATURE_V8) ? -4 : -3; + + if (s->vectors[ARMV7M_EXCP_RESET].prio != resetprio || s->vectors[ARMV7M_EXCP_NMI].prio != -2 || s->vectors[ARMV7M_EXCP_HARD].prio != -1) { return 1; @@ -XXX,XX +XXX,XX @@ static int nvic_security_post_load(void *opaque, int version_id) int i; /* Check for out of range priority settings */ - if (s->sec_vectors[ARMV7M_EXCP_HARD].prio != -1) { + if (s->sec_vectors[ARMV7M_EXCP_HARD].prio != -1 + && s->sec_vectors[ARMV7M_EXCP_HARD].prio != -3) { + /* We can't cross-check against AIRCR.BFHFNMINS as we don't know + * if the CPU state has been migrated yet; a mismatch won't + * cause the emulation to blow up, though. + */ return 1; } for (i = ARMV7M_EXCP_MEM; i < ARRAY_SIZE(s->sec_vectors); i++) { @@ -XXX,XX +XXX,XX @@ static Property props_nvic[] = { static void armv7m_nvic_reset(DeviceState *dev) { + int resetprio; NVICState *s = NVIC(dev); s->vectors[ARMV7M_EXCP_NMI].enabled = 1; @@ -XXX,XX +XXX,XX @@ static void armv7m_nvic_reset(DeviceState *dev) s->vectors[ARMV7M_EXCP_PENDSV].enabled = 1; s->vectors[ARMV7M_EXCP_SYSTICK].enabled = 1; - s->vectors[ARMV7M_EXCP_RESET].prio = -3; + resetprio = arm_feature(&s->cpu->env, ARM_FEATURE_V8) ? -4 : -3; + s->vectors[ARMV7M_EXCP_RESET].prio = resetprio; s->vectors[ARMV7M_EXCP_NMI].prio = -2; s->vectors[ARMV7M_EXCP_HARD].prio = -1; -- 2.7.4 If AIRCR.BFHFNMINS is clear, then although NonSecure HardFault can still be pended via SHCSR.HARDFAULTPENDED it mustn't actually preempt execution. The simple way to achieve this is to clear the enable bit for it, since the enable bit isn't guest visible. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-15-git-send-email-peter.maydell@linaro.org --- hw/intc/armv7m_nvic.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ static void nvic_writel(NVICState *s, uint32_t offset, uint32_t value, (R_V7M_AIRCR_SYSRESETREQS_MASK | R_V7M_AIRCR_BFHFNMINS_MASK | R_V7M_AIRCR_PRIS_MASK); - /* BFHFNMINS changes the priority of Secure HardFault */ + /* BFHFNMINS changes the priority of Secure HardFault, and + * allows a pending Non-secure HardFault to preempt (which + * we implement by marking it enabled). + */ if (cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK) { s->sec_vectors[ARMV7M_EXCP_HARD].prio = -3; + s->vectors[ARMV7M_EXCP_HARD].enabled = 1; } else { s->sec_vectors[ARMV7M_EXCP_HARD].prio = -1; + s->vectors[ARMV7M_EXCP_HARD].enabled = 0; } } nvic_irq_update(s); @@ -XXX,XX +XXX,XX @@ static void armv7m_nvic_reset(DeviceState *dev) NVICState *s = NVIC(dev); s->vectors[ARMV7M_EXCP_NMI].enabled = 1; - s->vectors[ARMV7M_EXCP_HARD].enabled = 1; /* MEM, BUS, and USAGE are enabled through * the System Handler Control register */ @@ -XXX,XX +XXX,XX @@ static void armv7m_nvic_reset(DeviceState *dev) /* AIRCR.BFHFNMINS resets to 0 so Secure HF is priority -1 (R_CMTC) */ s->sec_vectors[ARMV7M_EXCP_HARD].prio = -1; + /* If AIRCR.BFHFNMINS is 0 then NS HF is (effectively) disabled */ + s->vectors[ARMV7M_EXCP_HARD].enabled = 0; + } else { + s->vectors[ARMV7M_EXCP_HARD].enabled = 1; } /* Strictly speaking the reset handler should be enabled. -- 2.7.4 Update nvic_exec_prio() to support the v8M changes: * BASEPRI, FAULTMASK and PRIMASK are all banked * AIRCR.PRIS can affect NS priorities * AIRCR.BFHFNMINS affects FAULTMASK behaviour These changes mean that it's no longer possible to definitely say that if FAULTMASK is set it overrides PRIMASK, and if PRIMASK is set it overrides BASEPRI (since if PRIMASK_NS is set and AIRCR.PRIS is set then whether that 0x80 priority should take effect or the priority in BASEPRI_S depends on the value of BASEPRI_S, for instance). So we switch to the same approach used by the pseudocode of working through BASEPRI, PRIMASK and FAULTMASK and overriding the previous values if needed. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-16-git-send-email-peter.maydell@linaro.org --- hw/intc/armv7m_nvic.c | 51 ++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 42 insertions(+), 9 deletions(-) diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ static void nvic_recompute_state(NVICState *s) static inline int nvic_exec_prio(NVICState *s) { CPUARMState *env = &s->cpu->env; - int running; + int running = NVIC_NOEXC_PRIO; - if (env->v7m.faultmask[env->v7m.secure]) { - running = -1; - } else if (env->v7m.primask[env->v7m.secure]) { + if (env->v7m.basepri[M_REG_NS] > 0) { + running = exc_group_prio(s, env->v7m.basepri[M_REG_NS], M_REG_NS); + } + + if (env->v7m.basepri[M_REG_S] > 0) { + int basepri = exc_group_prio(s, env->v7m.basepri[M_REG_S], M_REG_S); + if (running > basepri) { + running = basepri; + } + } + + if (env->v7m.primask[M_REG_NS]) { + if (env->v7m.aircr & R_V7M_AIRCR_PRIS_MASK) { + if (running > NVIC_NS_PRIO_LIMIT) { + running = NVIC_NS_PRIO_LIMIT; + } + } else { + running = 0; + } + } + + if (env->v7m.primask[M_REG_S]) { running = 0; - } else if (env->v7m.basepri[env->v7m.secure] > 0) { - running = env->v7m.basepri[env->v7m.secure] & - nvic_gprio_mask(s, env->v7m.secure); - } else { - running = NVIC_NOEXC_PRIO; /* lower than any possible priority */ } + + if (env->v7m.faultmask[M_REG_NS]) { + if (env->v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK) { + running = -1; + } else { + if (env->v7m.aircr & R_V7M_AIRCR_PRIS_MASK) { + if (running > NVIC_NS_PRIO_LIMIT) { + running = NVIC_NS_PRIO_LIMIT; + } + } else { + running = 0; + } + } + } + + if (env->v7m.faultmask[M_REG_S]) { + running = (env->v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK) ? -3 : -1; + } + /* consider priority of active handler */ return MIN(running, s->exception_prio); } -- 2.7.4 Now that we have a banked FAULTMASK register and banked exceptions, we can implement the correct check in cpu_mmu_index() for whether the MPU_CTRL.HFNMIENA bit's effect should apply. This bit causes handlers which have requested a negative execution priority to run with the MPU disabled. In v8M the test has to check this for the current security state and so takes account of banking. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-17-git-send-email-peter.maydell@linaro.org --- target/arm/cpu.h | 21 ++++++++++++++++----- hw/intc/armv7m_nvic.c | 29 +++++++++++++++++++++++++++++ 2 files changed, 45 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 @@ int armv7m_nvic_complete_irq(void *opaque, int irq); * (v8M ARM ARM I_PKLD.) */ int armv7m_nvic_raw_execution_priority(void *opaque); +/** + * armv7m_nvic_neg_prio_requested: return true if the requested execution + * priority is negative for the specified security state. + * @opaque: the NVIC + * @secure: the security state to test + * This corresponds to the pseudocode IsReqExecPriNeg(). + */ +#ifndef CONFIG_USER_ONLY +bool armv7m_nvic_neg_prio_requested(void *opaque, bool secure); +#else +static inline bool armv7m_nvic_neg_prio_requested(void *opaque, bool secure) +{ + return false; +} +#endif /* Interface for defining coprocessor registers. * Registers are defined in tables of arm_cp_reginfo structs @@ -XXX,XX +XXX,XX @@ static inline int cpu_mmu_index(CPUARMState *env, bool ifetch) if (arm_feature(env, ARM_FEATURE_M)) { ARMMMUIdx mmu_idx = el == 0 ? ARMMMUIdx_MUser : ARMMMUIdx_MPriv; - /* Execution priority is negative if FAULTMASK is set or - * we're in a HardFault or NMI handler. - */ - if ((env->v7m.exception > 0 && env->v7m.exception <= 3) - || env->v7m.faultmask[env->v7m.secure]) { + if (armv7m_nvic_neg_prio_requested(env->nvic, env->v7m.secure)) { mmu_idx = ARMMMUIdx_MNegPri; } diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ static inline int nvic_exec_prio(NVICState *s) return MIN(running, s->exception_prio); } +bool armv7m_nvic_neg_prio_requested(void *opaque, bool secure) +{ + /* Return true if the requested execution priority is negative + * for the specified security state, ie that security state + * has an active NMI or HardFault or has set its FAULTMASK. + * Note that this is not the same as whether the execution + * priority is actually negative (for instance AIRCR.PRIS may + * mean we don't allow FAULTMASK_NS to actually make the execution + * priority negative). Compare pseudocode IsReqExcPriNeg(). + */ + NVICState *s = opaque; + + if (s->cpu->env.v7m.faultmask[secure]) { + return true; + } + + if (secure ? s->sec_vectors[ARMV7M_EXCP_HARD].active : + s->vectors[ARMV7M_EXCP_HARD].active) { + return true; + } + + if (s->vectors[ARMV7M_EXCP_NMI].active && + exc_targets_secure(s, ARMV7M_EXCP_NMI) == secure) { + return true; + } + + return false; +} + bool armv7m_nvic_can_take_pending_exception(void *opaque) { NVICState *s = opaque; -- 2.7.4 The ICSR NVIC register is banked for v8M. This doesn't require any new state, but it does mean that some bits are controlled by BFHNFNMINS and some bits must work with the correct banked exception. There is also a new in v8M PENDNMICLR bit. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-18-git-send-email-peter.maydell@linaro.org --- hw/intc/armv7m_nvic.c | 45 ++++++++++++++++++++++++++++++++------------- 1 file changed, 32 insertions(+), 13 deletions(-) diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ static uint32_t nvic_readl(NVICState *s, uint32_t offset, MemTxAttrs attrs) } case 0xd00: /* CPUID Base. */ return cpu->midr; - case 0xd04: /* Interrupt Control State. */ + case 0xd04: /* Interrupt Control State (ICSR) */ /* VECTACTIVE */ val = cpu->env.v7m.exception; /* VECTPENDING */ @@ -XXX,XX +XXX,XX @@ static uint32_t nvic_readl(NVICState *s, uint32_t offset, MemTxAttrs attrs) if (nvic_rettobase(s)) { val |= (1 << 11); } - /* PENDSTSET */ - if (s->vectors[ARMV7M_EXCP_SYSTICK].pending) { - val |= (1 << 26); - } - /* PENDSVSET */ - if (s->vectors[ARMV7M_EXCP_PENDSV].pending) { - val |= (1 << 28); + if (attrs.secure) { + /* PENDSTSET */ + if (s->sec_vectors[ARMV7M_EXCP_SYSTICK].pending) { + val |= (1 << 26); + } + /* PENDSVSET */ + if (s->sec_vectors[ARMV7M_EXCP_PENDSV].pending) { + val |= (1 << 28); + } + } else { + /* PENDSTSET */ + if (s->vectors[ARMV7M_EXCP_SYSTICK].pending) { + val |= (1 << 26); + } + /* PENDSVSET */ + if (s->vectors[ARMV7M_EXCP_PENDSV].pending) { + val |= (1 << 28); + } } /* NMIPENDSET */ - if (s->vectors[ARMV7M_EXCP_NMI].pending) { + if ((cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK) && + s->vectors[ARMV7M_EXCP_NMI].pending) { val |= (1 << 31); } - /* ISRPREEMPT not implemented */ + /* ISRPREEMPT: RES0 when halting debug not implemented */ + /* STTNS: RES0 for the Main Extension */ return val; case 0xd08: /* Vector Table Offset. */ return cpu->env.v7m.vecbase[attrs.secure]; @@ -XXX,XX +XXX,XX @@ static void nvic_writel(NVICState *s, uint32_t offset, uint32_t value, nvic_irq_update(s); break; } - case 0xd04: /* Interrupt Control State. */ - if (value & (1 << 31)) { - armv7m_nvic_set_pending(s, ARMV7M_EXCP_NMI, false); + case 0xd04: /* Interrupt Control State (ICSR) */ + if (cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK) { + if (value & (1 << 31)) { + armv7m_nvic_set_pending(s, ARMV7M_EXCP_NMI, false); + } else if (value & (1 << 30) && + arm_feature(&cpu->env, ARM_FEATURE_V8)) { + /* PENDNMICLR didn't exist in v7M */ + armv7m_nvic_clear_pending(s, ARMV7M_EXCP_NMI, false); + } } if (value & (1 << 28)) { armv7m_nvic_set_pending(s, ARMV7M_EXCP_PENDSV, attrs.secure); -- 2.7.4 Handle banking of SHCSR: some register bits are banked between Secure and Non-Secure, and some are only accessible to Secure. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-19-git-send-email-peter.maydell@linaro.org --- hw/intc/armv7m_nvic.c | 221 ++++++++++++++++++++++++++++++++++++++------------ 1 file changed, 169 insertions(+), 52 deletions(-) diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ static uint32_t nvic_readl(NVICState *s, uint32_t offset, MemTxAttrs attrs) val = cpu->env.v7m.ccr[attrs.secure]; val |= cpu->env.v7m.ccr[M_REG_NS] & R_V7M_CCR_BFHFNMIGN_MASK; return val; - case 0xd24: /* System Handler Status. */ + case 0xd24: /* System Handler Control and State (SHCSR) */ val = 0; - if (s->vectors[ARMV7M_EXCP_MEM].active) { - val |= (1 << 0); - } - if (s->vectors[ARMV7M_EXCP_BUS].active) { - val |= (1 << 1); - } - if (s->vectors[ARMV7M_EXCP_USAGE].active) { - val |= (1 << 3); + if (attrs.secure) { + if (s->sec_vectors[ARMV7M_EXCP_MEM].active) { + val |= (1 << 0); + } + if (s->sec_vectors[ARMV7M_EXCP_HARD].active) { + val |= (1 << 2); + } + if (s->sec_vectors[ARMV7M_EXCP_USAGE].active) { + val |= (1 << 3); + } + if (s->sec_vectors[ARMV7M_EXCP_SVC].active) { + val |= (1 << 7); + } + if (s->sec_vectors[ARMV7M_EXCP_PENDSV].active) { + val |= (1 << 10); + } + if (s->sec_vectors[ARMV7M_EXCP_SYSTICK].active) { + val |= (1 << 11); + } + if (s->sec_vectors[ARMV7M_EXCP_USAGE].pending) { + val |= (1 << 12); + } + if (s->sec_vectors[ARMV7M_EXCP_MEM].pending) { + val |= (1 << 13); + } + if (s->sec_vectors[ARMV7M_EXCP_SVC].pending) { + val |= (1 << 15); + } + if (s->sec_vectors[ARMV7M_EXCP_MEM].enabled) { + val |= (1 << 16); + } + if (s->sec_vectors[ARMV7M_EXCP_USAGE].enabled) { + val |= (1 << 18); + } + if (s->sec_vectors[ARMV7M_EXCP_HARD].pending) { + val |= (1 << 21); + } + /* SecureFault is not banked but is always RAZ/WI to NS */ + if (s->vectors[ARMV7M_EXCP_SECURE].active) { + val |= (1 << 4); + } + if (s->vectors[ARMV7M_EXCP_SECURE].enabled) { + val |= (1 << 19); + } + if (s->vectors[ARMV7M_EXCP_SECURE].pending) { + val |= (1 << 20); + } + } else { + if (s->vectors[ARMV7M_EXCP_MEM].active) { + val |= (1 << 0); + } + if (arm_feature(&cpu->env, ARM_FEATURE_V8)) { + /* HARDFAULTACT, HARDFAULTPENDED not present in v7M */ + if (s->vectors[ARMV7M_EXCP_HARD].active) { + val |= (1 << 2); + } + if (s->vectors[ARMV7M_EXCP_HARD].pending) { + val |= (1 << 21); + } + } + if (s->vectors[ARMV7M_EXCP_USAGE].active) { + val |= (1 << 3); + } + if (s->vectors[ARMV7M_EXCP_SVC].active) { + val |= (1 << 7); + } + if (s->vectors[ARMV7M_EXCP_PENDSV].active) { + val |= (1 << 10); + } + if (s->vectors[ARMV7M_EXCP_SYSTICK].active) { + val |= (1 << 11); + } + if (s->vectors[ARMV7M_EXCP_USAGE].pending) { + val |= (1 << 12); + } + if (s->vectors[ARMV7M_EXCP_MEM].pending) { + val |= (1 << 13); + } + if (s->vectors[ARMV7M_EXCP_SVC].pending) { + val |= (1 << 15); + } + if (s->vectors[ARMV7M_EXCP_MEM].enabled) { + val |= (1 << 16); + } + if (s->vectors[ARMV7M_EXCP_USAGE].enabled) { + val |= (1 << 18); + } } - if (s->vectors[ARMV7M_EXCP_SVC].active) { - val |= (1 << 7); + if (attrs.secure || (cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK)) { + if (s->vectors[ARMV7M_EXCP_BUS].active) { + val |= (1 << 1); + } + if (s->vectors[ARMV7M_EXCP_BUS].pending) { + val |= (1 << 14); + } + if (s->vectors[ARMV7M_EXCP_BUS].enabled) { + val |= (1 << 17); + } + if (arm_feature(&cpu->env, ARM_FEATURE_V8) && + s->vectors[ARMV7M_EXCP_NMI].active) { + /* NMIACT is not present in v7M */ + val |= (1 << 5); + } } + + /* TODO: this is RAZ/WI from NS if DEMCR.SDME is set */ if (s->vectors[ARMV7M_EXCP_DEBUG].active) { val |= (1 << 8); } - if (s->vectors[ARMV7M_EXCP_PENDSV].active) { - val |= (1 << 10); - } - if (s->vectors[ARMV7M_EXCP_SYSTICK].active) { - val |= (1 << 11); - } - if (s->vectors[ARMV7M_EXCP_USAGE].pending) { - val |= (1 << 12); - } - if (s->vectors[ARMV7M_EXCP_MEM].pending) { - val |= (1 << 13); - } - if (s->vectors[ARMV7M_EXCP_BUS].pending) { - val |= (1 << 14); - } - if (s->vectors[ARMV7M_EXCP_SVC].pending) { - val |= (1 << 15); - } - if (s->vectors[ARMV7M_EXCP_MEM].enabled) { - val |= (1 << 16); - } - if (s->vectors[ARMV7M_EXCP_BUS].enabled) { - val |= (1 << 17); - } - if (s->vectors[ARMV7M_EXCP_USAGE].enabled) { - val |= (1 << 18); - } return val; case 0xd28: /* Configurable Fault Status. */ /* The BFSR bits [15:8] are shared between security states @@ -XXX,XX +XXX,XX @@ static void nvic_writel(NVICState *s, uint32_t offset, uint32_t value, cpu->env.v7m.ccr[attrs.secure] = value; break; - case 0xd24: /* System Handler Control. */ - s->vectors[ARMV7M_EXCP_MEM].active = (value & (1 << 0)) != 0; - s->vectors[ARMV7M_EXCP_BUS].active = (value & (1 << 1)) != 0; - s->vectors[ARMV7M_EXCP_USAGE].active = (value & (1 << 3)) != 0; - s->vectors[ARMV7M_EXCP_SVC].active = (value & (1 << 7)) != 0; + case 0xd24: /* System Handler Control and State (SHCSR) */ + if (attrs.secure) { + s->sec_vectors[ARMV7M_EXCP_MEM].active = (value & (1 << 0)) != 0; + /* Secure HardFault active bit cannot be written */ + s->sec_vectors[ARMV7M_EXCP_USAGE].active = (value & (1 << 3)) != 0; + s->sec_vectors[ARMV7M_EXCP_SVC].active = (value & (1 << 7)) != 0; + s->sec_vectors[ARMV7M_EXCP_PENDSV].active = + (value & (1 << 10)) != 0; + s->sec_vectors[ARMV7M_EXCP_SYSTICK].active = + (value & (1 << 11)) != 0; + s->sec_vectors[ARMV7M_EXCP_USAGE].pending = + (value & (1 << 12)) != 0; + s->sec_vectors[ARMV7M_EXCP_MEM].pending = (value & (1 << 13)) != 0; + s->sec_vectors[ARMV7M_EXCP_SVC].pending = (value & (1 << 15)) != 0; + s->sec_vectors[ARMV7M_EXCP_MEM].enabled = (value & (1 << 16)) != 0; + s->sec_vectors[ARMV7M_EXCP_BUS].enabled = (value & (1 << 17)) != 0; + s->sec_vectors[ARMV7M_EXCP_USAGE].enabled = + (value & (1 << 18)) != 0; + /* SecureFault not banked, but RAZ/WI to NS */ + s->vectors[ARMV7M_EXCP_SECURE].active = (value & (1 << 4)) != 0; + s->vectors[ARMV7M_EXCP_SECURE].enabled = (value & (1 << 19)) != 0; + s->vectors[ARMV7M_EXCP_SECURE].pending = (value & (1 << 20)) != 0; + } else { + s->vectors[ARMV7M_EXCP_MEM].active = (value & (1 << 0)) != 0; + if (arm_feature(&cpu->env, ARM_FEATURE_V8)) { + /* HARDFAULTPENDED is not present in v7M */ + s->vectors[ARMV7M_EXCP_HARD].pending = (value & (1 << 21)) != 0; + } + s->vectors[ARMV7M_EXCP_USAGE].active = (value & (1 << 3)) != 0; + s->vectors[ARMV7M_EXCP_SVC].active = (value & (1 << 7)) != 0; + s->vectors[ARMV7M_EXCP_PENDSV].active = (value & (1 << 10)) != 0; + s->vectors[ARMV7M_EXCP_SYSTICK].active = (value & (1 << 11)) != 0; + s->vectors[ARMV7M_EXCP_USAGE].pending = (value & (1 << 12)) != 0; + s->vectors[ARMV7M_EXCP_MEM].pending = (value & (1 << 13)) != 0; + s->vectors[ARMV7M_EXCP_SVC].pending = (value & (1 << 15)) != 0; + s->vectors[ARMV7M_EXCP_MEM].enabled = (value & (1 << 16)) != 0; + s->vectors[ARMV7M_EXCP_USAGE].enabled = (value & (1 << 18)) != 0; + } + if (attrs.secure || (cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK)) { + s->vectors[ARMV7M_EXCP_BUS].active = (value & (1 << 1)) != 0; + s->vectors[ARMV7M_EXCP_BUS].pending = (value & (1 << 14)) != 0; + s->vectors[ARMV7M_EXCP_BUS].enabled = (value & (1 << 17)) != 0; + } + /* NMIACT can only be written if the write is of a zero, with + * BFHFNMINS 1, and by the CPU in secure state via the NS alias. + */ + if (!attrs.secure && cpu->env.v7m.secure && + (cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK) && + (value & (1 << 5)) == 0) { + s->vectors[ARMV7M_EXCP_NMI].active = 0; + } + /* HARDFAULTACT can only be written if the write is of a zero + * to the non-secure HardFault state by the CPU in secure state. + * The only case where we can be targeting the non-secure HF state + * when in secure state is if this is a write via the NS alias + * and BFHFNMINS is 1. + */ + if (!attrs.secure && cpu->env.v7m.secure && + (cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK) && + (value & (1 << 2)) == 0) { + s->vectors[ARMV7M_EXCP_HARD].active = 0; + } + + /* TODO: this is RAZ/WI from NS if DEMCR.SDME is set */ s->vectors[ARMV7M_EXCP_DEBUG].active = (value & (1 << 8)) != 0; - s->vectors[ARMV7M_EXCP_PENDSV].active = (value & (1 << 10)) != 0; - s->vectors[ARMV7M_EXCP_SYSTICK].active = (value & (1 << 11)) != 0; - s->vectors[ARMV7M_EXCP_USAGE].pending = (value & (1 << 12)) != 0; - s->vectors[ARMV7M_EXCP_MEM].pending = (value & (1 << 13)) != 0; - s->vectors[ARMV7M_EXCP_BUS].pending = (value & (1 << 14)) != 0; - s->vectors[ARMV7M_EXCP_SVC].pending = (value & (1 << 15)) != 0; - s->vectors[ARMV7M_EXCP_MEM].enabled = (value & (1 << 16)) != 0; - s->vectors[ARMV7M_EXCP_BUS].enabled = (value & (1 << 17)) != 0; - s->vectors[ARMV7M_EXCP_USAGE].enabled = (value & (1 << 18)) != 0; nvic_irq_update(s); break; case 0xd28: /* Configurable Fault Status. */ -- 2.7.4 Update armv7m_nvic_acknowledge_irq() and armv7m_nvic_complete_irq() to handle banked exceptions: * acknowledge needs to use the correct vector, which may be in sec_vectors[] * acknowledge needs to return to its caller whether the exception should be taken to secure or non-secure state * complete needs its caller to tell it whether the exception being completed is a secure one or not Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505240046-11454-20-git-send-email-peter.maydell@linaro.org --- target/arm/cpu.h | 15 +++++++++++++-- hw/intc/armv7m_nvic.c | 26 ++++++++++++++++++++------ target/arm/helper.c | 8 +++++--- hw/intc/trace-events | 4 ++-- 4 files changed, 40 insertions(+), 13 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 @@ static inline bool armv7m_nvic_can_take_pending_exception(void *opaque) * of architecturally banked exceptions. */ void armv7m_nvic_set_pending(void *opaque, int irq, bool secure); -void armv7m_nvic_acknowledge_irq(void *opaque); +/** + * armv7m_nvic_acknowledge_irq: make highest priority pending exception active + * @opaque: the NVIC + * + * Move the current highest priority pending exception from the pending + * state to the active state, and update v7m.exception to indicate that + * it is the exception currently being handled. + * + * Returns: true if exception should be taken to Secure state, false for NS + */ +bool armv7m_nvic_acknowledge_irq(void *opaque); /** * armv7m_nvic_complete_irq: complete specified interrupt or exception * @opaque: the NVIC * @irq: the exception number to complete + * @secure: true if this exception was secure * * Returns: -1 if the irq was not active * 1 if completing this irq brought us back to base (no active irqs) * 0 if there is still an irq active after this one was completed * (Ignoring -1, this is the same as the RETTOBASE value before completion.) */ -int armv7m_nvic_complete_irq(void *opaque, int irq); +int armv7m_nvic_complete_irq(void *opaque, int irq, bool secure); /** * armv7m_nvic_raw_execution_priority: return the raw execution priority * @opaque: the NVIC diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/armv7m_nvic.c +++ b/hw/intc/armv7m_nvic.c @@ -XXX,XX +XXX,XX @@ void armv7m_nvic_set_pending(void *opaque, int irq, bool secure) } /* Make pending IRQ active. */ -void armv7m_nvic_acknowledge_irq(void *opaque) +bool armv7m_nvic_acknowledge_irq(void *opaque) { NVICState *s = (NVICState *)opaque; CPUARMState *env = &s->cpu->env; const int pending = s->vectpending; const int running = nvic_exec_prio(s); VecInfo *vec; + bool targets_secure; assert(pending > ARMV7M_EXCP_RESET && pending < s->num_irq); - vec = &s->vectors[pending]; + if (s->vectpending_is_s_banked) { + vec = &s->sec_vectors[pending]; + targets_secure = true; + } else { + vec = &s->vectors[pending]; + targets_secure = !exc_is_banked(s->vectpending) && + exc_targets_secure(s, s->vectpending); + } assert(vec->enabled); assert(vec->pending); assert(s->vectpending_prio < running); - trace_nvic_acknowledge_irq(pending, s->vectpending_prio); + trace_nvic_acknowledge_irq(pending, s->vectpending_prio, targets_secure); vec->active = 1; vec->pending = 0; @@ -XXX,XX +XXX,XX @@ void armv7m_nvic_acknowledge_irq(void *opaque) env->v7m.exception = s->vectpending; nvic_irq_update(s); + + return targets_secure; } -int armv7m_nvic_complete_irq(void *opaque, int irq) +int armv7m_nvic_complete_irq(void *opaque, int irq, bool secure) { NVICState *s = (NVICState *)opaque; VecInfo *vec; @@ -XXX,XX +XXX,XX @@ int armv7m_nvic_complete_irq(void *opaque, int irq) assert(irq > ARMV7M_EXCP_RESET && irq < s->num_irq); - vec = &s->vectors[irq]; + if (secure && exc_is_banked(irq)) { + vec = &s->sec_vectors[irq]; + } else { + vec = &s->vectors[irq]; + } - trace_nvic_complete_irq(irq); + trace_nvic_complete_irq(irq, secure); if (!vec->active) { /* Tell the caller this was an illegal exception return */ 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) bool return_to_sp_process = false; bool return_to_handler = false; bool rettobase = false; + bool exc_secure = false; /* We can only get here from an EXCP_EXCEPTION_EXIT, and * gen_bx_excret() enforces the architectural rule @@ -XXX,XX +XXX,XX @@ static void do_v7m_exception_exit(ARMCPU *cpu) * which security state's faultmask to clear. (v8M ARM ARM R_KBNF.) */ if (arm_feature(env, ARM_FEATURE_M_SECURITY)) { - int es = excret & R_V7M_EXCRET_ES_MASK; + exc_secure = excret & R_V7M_EXCRET_ES_MASK; if (armv7m_nvic_raw_execution_priority(env->nvic) >= 0) { - env->v7m.faultmask[es] = 0; + env->v7m.faultmask[exc_secure] = 0; } } else { env->v7m.faultmask[M_REG_NS] = 0; } } - switch (armv7m_nvic_complete_irq(env->nvic, env->v7m.exception)) { + switch (armv7m_nvic_complete_irq(env->nvic, env->v7m.exception, + exc_secure)) { case -1: /* attempt to exit an exception that isn't active */ ufault = true; diff --git a/hw/intc/trace-events b/hw/intc/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/trace-events +++ b/hw/intc/trace-events @@ -XXX,XX +XXX,XX @@ nvic_escalate_disabled(int irq) "NVIC escalating irq %d to HardFault: disabled" nvic_set_pending(int irq, bool secure, int en, int prio) "NVIC set pending irq %d secure-bank %d (enabled: %d priority %d)" nvic_clear_pending(int irq, bool secure, int en, int prio) "NVIC clear pending irq %d secure-bank %d (enabled: %d priority %d)" nvic_set_pending_level(int irq) "NVIC set pending: irq %d higher prio than vectpending: setting irq line to 1" -nvic_acknowledge_irq(int irq, int prio) "NVIC acknowledge IRQ: %d now active (prio %d)" -nvic_complete_irq(int irq) "NVIC complete IRQ %d" +nvic_acknowledge_irq(int irq, int prio, bool targets_secure) "NVIC acknowledge IRQ: %d now active (prio %d targets_secure %d)" +nvic_complete_irq(int irq, bool secure) "NVIC complete IRQ %d (secure %d)" nvic_set_irq_level(int irq, int level) "NVIC external irq %d level set to %d" nvic_sysreg_read(uint64_t addr, uint32_t value, unsigned size) "NVIC sysreg read addr 0x%" PRIx64 " data 0x%" PRIx32 " size %u" nvic_sysreg_write(uint64_t addr, uint32_t value, unsigned size) "NVIC sysreg write addr 0x%" PRIx64 " data 0x%" PRIx32 " size %u" -- 2.7.4 In the A64 decoder, we have a lot of references to section numbers from version A.a of the v8A ARM ARM (DDI0487). This version of the document is now long obsolete (we are currently on revision B.a), and various intervening versions renumbered all the sections. The most recent B.a version of the document doesn't assign section numbers at all to the individual instruction classes in the way that the various A.x versions did. The simplest thing to do is just to delete all the out of date C.x.x references. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Message-id: 20170915150849.23557-1-peter.maydell@linaro.org --- target/arm/translate-a64.c | 227 +++++++++++++++++++++++---------------------- 1 file changed, 114 insertions(+), 113 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static inline AArch64DecodeFn *lookup_disas_fn(const AArch64DecodeTable *table, } /* - * the instruction disassembly implemented here matches - * the instruction encoding classifications in chapter 3 (C3) - * of the ARM Architecture Reference Manual (DDI0487A_a) + * The instruction disassembly implemented here matches + * the instruction encoding classifications in chapter C4 + * of the ARM Architecture Reference Manual (DDI0487B_a); + * classification names and decode diagrams here should generally + * match up with those in the manual. */ -/* C3.2.7 Unconditional branch (immediate) +/* Unconditional branch (immediate) * 31 30 26 25 0 * +----+-----------+-------------------------------------+ * | op | 0 0 1 0 1 | imm26 | @@ -XXX,XX +XXX,XX @@ static void disas_uncond_b_imm(DisasContext *s, uint32_t insn) uint64_t addr = s->pc + sextract32(insn, 0, 26) * 4 - 4; if (insn & (1U << 31)) { - /* C5.6.26 BL Branch with link */ + /* BL Branch with link */ tcg_gen_movi_i64(cpu_reg(s, 30), s->pc); } - /* C5.6.20 B Branch / C5.6.26 BL Branch with link */ + /* B Branch / BL Branch with link */ gen_goto_tb(s, 0, addr); } -/* C3.2.1 Compare & branch (immediate) +/* Compare and branch (immediate) * 31 30 25 24 23 5 4 0 * +----+-------------+----+---------------------+--------+ * | sf | 0 1 1 0 1 0 | op | imm19 | Rt | @@ -XXX,XX +XXX,XX @@ static void disas_comp_b_imm(DisasContext *s, uint32_t insn) gen_goto_tb(s, 1, addr); } -/* C3.2.5 Test & branch (immediate) +/* Test and branch (immediate) * 31 30 25 24 23 19 18 5 4 0 * +----+-------------+----+-------+-------------+------+ * | b5 | 0 1 1 0 1 1 | op | b40 | imm14 | Rt | @@ -XXX,XX +XXX,XX @@ static void disas_test_b_imm(DisasContext *s, uint32_t insn) gen_goto_tb(s, 1, addr); } -/* C3.2.2 / C5.6.19 Conditional branch (immediate) +/* Conditional branch (immediate) * 31 25 24 23 5 4 3 0 * +---------------+----+---------------------+----+------+ * | 0 1 0 1 0 1 0 | o1 | imm19 | o0 | cond | @@ -XXX,XX +XXX,XX @@ static void disas_cond_b_imm(DisasContext *s, uint32_t insn) } } -/* C5.6.68 HINT */ +/* HINT instruction group, including various allocated HINTs */ static void handle_hint(DisasContext *s, uint32_t insn, unsigned int op1, unsigned int op2, unsigned int crm) { @@ -XXX,XX +XXX,XX @@ static void handle_sync(DisasContext *s, uint32_t insn, } } -/* C5.6.130 MSR (immediate) - move immediate to processor state field */ +/* MSR (immediate) - move immediate to processor state field */ static void handle_msr_i(DisasContext *s, uint32_t insn, unsigned int op1, unsigned int op2, unsigned int crm) { @@ -XXX,XX +XXX,XX @@ static void gen_set_nzcv(TCGv_i64 tcg_rt) tcg_temp_free_i32(nzcv); } -/* C5.6.129 MRS - move from system register - * C5.6.131 MSR (register) - move to system register - * C5.6.204 SYS - * C5.6.205 SYSL +/* MRS - move from system register + * MSR (register) - move to system register + * SYS + * SYSL * These are all essentially the same insn in 'read' and 'write' * versions, with varying op0 fields. */ @@ -XXX,XX +XXX,XX @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread, } } -/* C3.2.4 System +/* System * 31 22 21 20 19 18 16 15 12 11 8 7 5 4 0 * +---------------------+---+-----+-----+-------+-------+-----+------+ * | 1 1 0 1 0 1 0 1 0 0 | L | op0 | op1 | CRn | CRm | op2 | Rt | @@ -XXX,XX +XXX,XX @@ static void disas_system(DisasContext *s, uint32_t insn) return; } switch (crn) { - case 2: /* C5.6.68 HINT */ + case 2: /* HINT (including allocated hints like NOP, YIELD, etc) */ handle_hint(s, insn, op1, op2, crm); break; case 3: /* CLREX, DSB, DMB, ISB */ handle_sync(s, insn, op1, op2, crm); break; - case 4: /* C5.6.130 MSR (immediate) */ + case 4: /* MSR (immediate) */ handle_msr_i(s, insn, op1, op2, crm); break; default: @@ -XXX,XX +XXX,XX @@ static void disas_system(DisasContext *s, uint32_t insn) handle_sys(s, insn, l, op0, op1, op2, crn, crm, rt); } -/* C3.2.3 Exception generation +/* Exception generation * * 31 24 23 21 20 5 4 2 1 0 * +-----------------+-----+------------------------+-----+----+ @@ -XXX,XX +XXX,XX @@ static void disas_exc(DisasContext *s, uint32_t insn) } } -/* C3.2.7 Unconditional branch (register) +/* Unconditional branch (register) * 31 25 24 21 20 16 15 10 9 5 4 0 * +---------------+-------+-------+-------+------+-------+ * | 1 1 0 1 0 1 1 | opc | op2 | op3 | Rn | op4 | @@ -XXX,XX +XXX,XX @@ static void disas_uncond_b_reg(DisasContext *s, uint32_t insn) s->base.is_jmp = DISAS_JUMP; } -/* C3.2 Branches, exception generating and system instructions */ +/* Branches, exception generating and system instructions */ static void disas_b_exc_sys(DisasContext *s, uint32_t insn) { switch (extract32(insn, 25, 7)) { @@ -XXX,XX +XXX,XX @@ static bool disas_ldst_compute_iss_sf(int size, bool is_signed, int opc) return regsize == 64; } -/* C3.3.6 Load/store exclusive +/* Load/store exclusive * * 31 30 29 24 23 22 21 20 16 15 14 10 9 5 4 0 * +-----+-------------+----+---+----+------+----+-------+------+------+ @@ -XXX,XX +XXX,XX @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) } /* - * C3.3.5 Load register (literal) + * Load register (literal) * * 31 30 29 27 26 25 24 23 5 4 0 * +-----+-------+---+-----+-------------------+-------+ @@ -XXX,XX +XXX,XX @@ static void disas_ld_lit(DisasContext *s, uint32_t insn) } /* - * C5.6.80 LDNP (Load Pair - non-temporal hint) - * C5.6.81 LDP (Load Pair - non vector) - * C5.6.82 LDPSW (Load Pair Signed Word - non vector) - * C5.6.176 STNP (Store Pair - non-temporal hint) - * C5.6.177 STP (Store Pair - non vector) - * C6.3.165 LDNP (Load Pair of SIMD&FP - non-temporal hint) - * C6.3.165 LDP (Load Pair of SIMD&FP) - * C6.3.284 STNP (Store Pair of SIMD&FP - non-temporal hint) - * C6.3.284 STP (Store Pair of SIMD&FP) + * LDNP (Load Pair - non-temporal hint) + * LDP (Load Pair - non vector) + * LDPSW (Load Pair Signed Word - non vector) + * STNP (Store Pair - non-temporal hint) + * STP (Store Pair - non vector) + * LDNP (Load Pair of SIMD&FP - non-temporal hint) + * LDP (Load Pair of SIMD&FP) + * STNP (Store Pair of SIMD&FP - non-temporal hint) + * STP (Store Pair of SIMD&FP) * * 31 30 29 27 26 25 24 23 22 21 15 14 10 9 5 4 0 * +-----+-------+---+---+-------+---+-----------------------------+ @@ -XXX,XX +XXX,XX @@ static void disas_ldst_pair(DisasContext *s, uint32_t insn) } /* - * C3.3.8 Load/store (immediate post-indexed) - * C3.3.9 Load/store (immediate pre-indexed) - * C3.3.12 Load/store (unscaled immediate) + * Load/store (immediate post-indexed) + * Load/store (immediate pre-indexed) + * Load/store (unscaled immediate) * * 31 30 29 27 26 25 24 23 22 21 20 12 11 10 9 5 4 0 * +----+-------+---+-----+-----+---+--------+-----+------+------+ @@ -XXX,XX +XXX,XX @@ static void disas_ldst_reg_imm9(DisasContext *s, uint32_t insn, } /* - * C3.3.10 Load/store (register offset) + * Load/store (register offset) * * 31 30 29 27 26 25 24 23 22 21 20 16 15 13 12 11 10 9 5 4 0 * +----+-------+---+-----+-----+---+------+-----+--+-----+----+----+ @@ -XXX,XX +XXX,XX @@ static void disas_ldst_reg_roffset(DisasContext *s, uint32_t insn, } /* - * C3.3.13 Load/store (unsigned immediate) + * Load/store (unsigned immediate) * * 31 30 29 27 26 25 24 23 22 21 10 9 5 * +----+-------+---+-----+-----+------------+-------+------+ @@ -XXX,XX +XXX,XX @@ static void disas_ldst_reg(DisasContext *s, uint32_t insn) } } -/* C3.3.1 AdvSIMD load/store multiple structures +/* AdvSIMD load/store multiple structures * * 31 30 29 23 22 21 16 15 12 11 10 9 5 4 0 * +---+---+---------------+---+-------------+--------+------+------+------+ * | 0 | Q | 0 0 1 1 0 0 0 | L | 0 0 0 0 0 0 | opcode | size | Rn | Rt | * +---+---+---------------+---+-------------+--------+------+------+------+ * - * C3.3.2 AdvSIMD load/store multiple structures (post-indexed) + * AdvSIMD load/store multiple structures (post-indexed) * * 31 30 29 23 22 21 20 16 15 12 11 10 9 5 4 0 * +---+---+---------------+---+---+---------+--------+------+------+------+ @@ -XXX,XX +XXX,XX @@ static void disas_ldst_multiple_struct(DisasContext *s, uint32_t insn) tcg_temp_free_i64(tcg_addr); } -/* C3.3.3 AdvSIMD load/store single structure +/* AdvSIMD load/store single structure * * 31 30 29 23 22 21 20 16 15 13 12 11 10 9 5 4 0 * +---+---+---------------+-----+-----------+-----+---+------+------+------+ * | 0 | Q | 0 0 1 1 0 1 0 | L R | 0 0 0 0 0 | opc | S | size | Rn | Rt | * +---+---+---------------+-----+-----------+-----+---+------+------+------+ * - * C3.3.4 AdvSIMD load/store single structure (post-indexed) + * AdvSIMD load/store single structure (post-indexed) * * 31 30 29 23 22 21 20 16 15 13 12 11 10 9 5 4 0 * +---+---+---------------+-----+-----------+-----+---+------+------+------+ @@ -XXX,XX +XXX,XX @@ static void disas_ldst_single_struct(DisasContext *s, uint32_t insn) tcg_temp_free_i64(tcg_addr); } -/* C3.3 Loads and stores */ +/* Loads and stores */ static void disas_ldst(DisasContext *s, uint32_t insn) { switch (extract32(insn, 24, 6)) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst(DisasContext *s, uint32_t insn) } } -/* C3.4.6 PC-rel. addressing +/* PC-rel. addressing * 31 30 29 28 24 23 5 4 0 * +----+-------+-----------+-------------------+------+ * | op | immlo | 1 0 0 0 0 | immhi | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_pc_rel_adr(DisasContext *s, uint32_t insn) } /* - * C3.4.1 Add/subtract (immediate) + * Add/subtract (immediate) * * 31 30 29 28 24 23 22 21 10 9 5 4 0 * +--+--+--+-----------+-----+-------------+-----+-----+ @@ -XXX,XX +XXX,XX @@ static bool logic_imm_decode_wmask(uint64_t *result, unsigned int immn, return true; } -/* C3.4.4 Logical (immediate) +/* Logical (immediate) * 31 30 29 28 23 22 21 16 15 10 9 5 4 0 * +----+-----+-------------+---+------+------+------+------+ * | sf | opc | 1 0 0 1 0 0 | N | immr | imms | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_logic_imm(DisasContext *s, uint32_t insn) } /* - * C3.4.5 Move wide (immediate) + * Move wide (immediate) * * 31 30 29 28 23 22 21 20 5 4 0 * +--+-----+-------------+-----+----------------+------+ @@ -XXX,XX +XXX,XX @@ static void disas_movw_imm(DisasContext *s, uint32_t insn) } } -/* C3.4.2 Bitfield +/* Bitfield * 31 30 29 28 23 22 21 16 15 10 9 5 4 0 * +----+-----+-------------+---+------+------+------+------+ * | sf | opc | 1 0 0 1 1 0 | N | immr | imms | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_bitfield(DisasContext *s, uint32_t insn) } } -/* C3.4.3 Extract +/* Extract * 31 30 29 28 23 22 21 20 16 15 10 9 5 4 0 * +----+------+-------------+---+----+------+--------+------+------+ * | sf | op21 | 1 0 0 1 1 1 | N | o0 | Rm | imms | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_extract(DisasContext *s, uint32_t insn) } } -/* C3.4 Data processing - immediate */ +/* Data processing - immediate */ static void disas_data_proc_imm(DisasContext *s, uint32_t insn) { switch (extract32(insn, 23, 6)) { @@ -XXX,XX +XXX,XX @@ static void shift_reg_imm(TCGv_i64 dst, TCGv_i64 src, int sf, } } -/* C3.5.10 Logical (shifted register) +/* Logical (shifted register) * 31 30 29 28 24 23 22 21 20 16 15 10 9 5 4 0 * +----+-----+-----------+-------+---+------+--------+------+------+ * | sf | opc | 0 1 0 1 0 | shift | N | Rm | imm6 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_logic_reg(DisasContext *s, uint32_t insn) } /* - * C3.5.1 Add/subtract (extended register) + * Add/subtract (extended register) * * 31|30|29|28 24|23 22|21|20 16|15 13|12 10|9 5|4 0| * +--+--+--+-----------+-----+--+-------+------+------+----+----+ @@ -XXX,XX +XXX,XX @@ static void disas_add_sub_ext_reg(DisasContext *s, uint32_t insn) } /* - * C3.5.2 Add/subtract (shifted register) + * Add/subtract (shifted register) * * 31 30 29 28 24 23 22 21 20 16 15 10 9 5 4 0 * +--+--+--+-----------+-----+--+-------+---------+------+------+ @@ -XXX,XX +XXX,XX @@ static void disas_add_sub_reg(DisasContext *s, uint32_t insn) tcg_temp_free_i64(tcg_result); } -/* C3.5.9 Data-processing (3 source) - - 31 30 29 28 24 23 21 20 16 15 14 10 9 5 4 0 - +--+------+-----------+------+------+----+------+------+------+ - |sf| op54 | 1 1 0 1 1 | op31 | Rm | o0 | Ra | Rn | Rd | - +--+------+-----------+------+------+----+------+------+------+ - +/* Data-processing (3 source) + * + * 31 30 29 28 24 23 21 20 16 15 14 10 9 5 4 0 + * +--+------+-----------+------+------+----+------+------+------+ + * |sf| op54 | 1 1 0 1 1 | op31 | Rm | o0 | Ra | Rn | Rd | + * +--+------+-----------+------+------+----+------+------+------+ */ static void disas_data_proc_3src(DisasContext *s, uint32_t insn) { @@ -XXX,XX +XXX,XX @@ static void disas_data_proc_3src(DisasContext *s, uint32_t insn) tcg_temp_free_i64(tcg_tmp); } -/* C3.5.3 - Add/subtract (with carry) +/* Add/subtract (with carry) * 31 30 29 28 27 26 25 24 23 22 21 20 16 15 10 9 5 4 0 * +--+--+--+------------------------+------+---------+------+-----+ * |sf|op| S| 1 1 0 1 0 0 0 0 | rm | opcode2 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_adc_sbc(DisasContext *s, uint32_t insn) } } -/* C3.5.4 - C3.5.5 Conditional compare (immediate / register) +/* Conditional compare (immediate / register) * 31 30 29 28 27 26 25 24 23 22 21 20 16 15 12 11 10 9 5 4 3 0 * +--+--+--+------------------------+--------+------+----+--+------+--+-----+ * |sf|op| S| 1 1 0 1 0 0 1 0 |imm5/rm | cond |i/r |o2| Rn |o3|nzcv | @@ -XXX,XX +XXX,XX @@ static void disas_cc(DisasContext *s, uint32_t insn) tcg_temp_free_i32(tcg_t2); } -/* C3.5.6 Conditional select +/* Conditional select * 31 30 29 28 21 20 16 15 12 11 10 9 5 4 0 * +----+----+---+-----------------+------+------+-----+------+------+ * | sf | op | S | 1 1 0 1 0 1 0 0 | Rm | cond | op2 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void handle_rbit(DisasContext *s, unsigned int sf, } } -/* C5.6.149 REV with sf==1, opcode==3 ("REV64") */ +/* REV with sf==1, opcode==3 ("REV64") */ static void handle_rev64(DisasContext *s, unsigned int sf, unsigned int rn, unsigned int rd) { @@ -XXX,XX +XXX,XX @@ static void handle_rev64(DisasContext *s, unsigned int sf, tcg_gen_bswap64_i64(cpu_reg(s, rd), cpu_reg(s, rn)); } -/* C5.6.149 REV with sf==0, opcode==2 - * C5.6.151 REV32 (sf==1, opcode==2) +/* REV with sf==0, opcode==2 + * REV32 (sf==1, opcode==2) */ static void handle_rev32(DisasContext *s, unsigned int sf, unsigned int rn, unsigned int rd) @@ -XXX,XX +XXX,XX @@ static void handle_rev32(DisasContext *s, unsigned int sf, } } -/* C5.6.150 REV16 (opcode==1) */ +/* REV16 (opcode==1) */ static void handle_rev16(DisasContext *s, unsigned int sf, unsigned int rn, unsigned int rd) { @@ -XXX,XX +XXX,XX @@ static void handle_rev16(DisasContext *s, unsigned int sf, tcg_temp_free_i64(tcg_tmp); } -/* C3.5.7 Data-processing (1 source) +/* Data-processing (1 source) * 31 30 29 28 21 20 16 15 10 9 5 4 0 * +----+---+---+-----------------+---------+--------+------+------+ * | sf | 1 | S | 1 1 0 1 0 1 1 0 | opcode2 | opcode | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void handle_div(DisasContext *s, bool is_signed, unsigned int sf, } } -/* C5.6.115 LSLV, C5.6.118 LSRV, C5.6.17 ASRV, C5.6.154 RORV */ +/* LSLV, LSRV, ASRV, RORV */ static void handle_shift_reg(DisasContext *s, enum a64_shift_type shift_type, unsigned int sf, unsigned int rm, unsigned int rn, unsigned int rd) @@ -XXX,XX +XXX,XX @@ static void handle_crc32(DisasContext *s, tcg_temp_free_i32(tcg_bytes); } -/* C3.5.8 Data-processing (2 source) +/* Data-processing (2 source) * 31 30 29 28 21 20 16 15 10 9 5 4 0 * +----+---+---+-----------------+------+--------+------+------+ * | sf | 0 | S | 1 1 0 1 0 1 1 0 | Rm | opcode | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_data_proc_2src(DisasContext *s, uint32_t insn) } } -/* C3.5 Data processing - register */ +/* Data processing - register */ static void disas_data_proc_reg(DisasContext *s, uint32_t insn) { switch (extract32(insn, 24, 5)) { @@ -XXX,XX +XXX,XX @@ static void handle_fp_compare(DisasContext *s, bool is_double, tcg_temp_free_i64(tcg_flags); } -/* C3.6.22 Floating point compare +/* Floating point compare * 31 30 29 28 24 23 22 21 20 16 15 14 13 10 9 5 4 0 * +---+---+---+-----------+------+---+------+-----+---------+------+-------+ * | M | 0 | S | 1 1 1 1 0 | type | 1 | Rm | op | 1 0 0 0 | Rn | op2 | @@ -XXX,XX +XXX,XX @@ static void disas_fp_compare(DisasContext *s, uint32_t insn) handle_fp_compare(s, type, rn, rm, opc & 1, opc & 2); } -/* C3.6.23 Floating point conditional compare +/* Floating point conditional compare * 31 30 29 28 24 23 22 21 20 16 15 12 11 10 9 5 4 3 0 * +---+---+---+-----------+------+---+------+------+-----+------+----+------+ * | M | 0 | S | 1 1 1 1 0 | type | 1 | Rm | cond | 0 1 | Rn | op | nzcv | @@ -XXX,XX +XXX,XX @@ static void disas_fp_ccomp(DisasContext *s, uint32_t insn) } } -/* C3.6.24 Floating point conditional select +/* Floating point conditional select * 31 30 29 28 24 23 22 21 20 16 15 12 11 10 9 5 4 0 * +---+---+---+-----------+------+---+------+------+-----+------+------+ * | M | 0 | S | 1 1 1 1 0 | type | 1 | Rm | cond | 1 1 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_fp_csel(DisasContext *s, uint32_t insn) tcg_temp_free_i64(t_true); } -/* C3.6.25 Floating-point data-processing (1 source) - single precision */ +/* Floating-point data-processing (1 source) - single precision */ static void handle_fp_1src_single(DisasContext *s, int opcode, int rd, int rn) { TCGv_ptr fpst; @@ -XXX,XX +XXX,XX @@ static void handle_fp_1src_single(DisasContext *s, int opcode, int rd, int rn) tcg_temp_free_i32(tcg_res); } -/* C3.6.25 Floating-point data-processing (1 source) - double precision */ +/* Floating-point data-processing (1 source) - double precision */ static void handle_fp_1src_double(DisasContext *s, int opcode, int rd, int rn) { TCGv_ptr fpst; @@ -XXX,XX +XXX,XX @@ static void handle_fp_fcvt(DisasContext *s, int opcode, } } -/* C3.6.25 Floating point data-processing (1 source) +/* Floating point data-processing (1 source) * 31 30 29 28 24 23 22 21 20 15 14 10 9 5 4 0 * +---+---+---+-----------+------+---+--------+-----------+------+------+ * | M | 0 | S | 1 1 1 1 0 | type | 1 | opcode | 1 0 0 0 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_fp_1src(DisasContext *s, uint32_t insn) } } -/* C3.6.26 Floating-point data-processing (2 source) - single precision */ +/* Floating-point data-processing (2 source) - single precision */ static void handle_fp_2src_single(DisasContext *s, int opcode, int rd, int rn, int rm) { @@ -XXX,XX +XXX,XX @@ static void handle_fp_2src_single(DisasContext *s, int opcode, tcg_temp_free_i32(tcg_res); } -/* C3.6.26 Floating-point data-processing (2 source) - double precision */ +/* Floating-point data-processing (2 source) - double precision */ static void handle_fp_2src_double(DisasContext *s, int opcode, int rd, int rn, int rm) { @@ -XXX,XX +XXX,XX @@ static void handle_fp_2src_double(DisasContext *s, int opcode, tcg_temp_free_i64(tcg_res); } -/* C3.6.26 Floating point data-processing (2 source) +/* Floating point data-processing (2 source) * 31 30 29 28 24 23 22 21 20 16 15 12 11 10 9 5 4 0 * +---+---+---+-----------+------+---+------+--------+-----+------+------+ * | M | 0 | S | 1 1 1 1 0 | type | 1 | Rm | opcode | 1 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_fp_2src(DisasContext *s, uint32_t insn) } } -/* C3.6.27 Floating-point data-processing (3 source) - single precision */ +/* Floating-point data-processing (3 source) - single precision */ static void handle_fp_3src_single(DisasContext *s, bool o0, bool o1, int rd, int rn, int rm, int ra) { @@ -XXX,XX +XXX,XX @@ static void handle_fp_3src_single(DisasContext *s, bool o0, bool o1, tcg_temp_free_i32(tcg_res); } -/* C3.6.27 Floating-point data-processing (3 source) - double precision */ +/* Floating-point data-processing (3 source) - double precision */ static void handle_fp_3src_double(DisasContext *s, bool o0, bool o1, int rd, int rn, int rm, int ra) { @@ -XXX,XX +XXX,XX @@ static void handle_fp_3src_double(DisasContext *s, bool o0, bool o1, tcg_temp_free_i64(tcg_res); } -/* C3.6.27 Floating point data-processing (3 source) +/* Floating point data-processing (3 source) * 31 30 29 28 24 23 22 21 20 16 15 14 10 9 5 4 0 * +---+---+---+-----------+------+----+------+----+------+------+------+ * | M | 0 | S | 1 1 1 1 1 | type | o1 | Rm | o0 | Ra | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_fp_3src(DisasContext *s, uint32_t insn) } } -/* C3.6.28 Floating point immediate +/* Floating point immediate * 31 30 29 28 24 23 22 21 20 13 12 10 9 5 4 0 * +---+---+---+-----------+------+---+------------+-------+------+------+ * | M | 0 | S | 1 1 1 1 0 | type | 1 | imm8 | 1 0 0 | imm5 | Rd | @@ -XXX,XX +XXX,XX @@ static void handle_fpfpcvt(DisasContext *s, int rd, int rn, int opcode, tcg_temp_free_i32(tcg_shift); } -/* C3.6.29 Floating point <-> fixed point conversions +/* Floating point <-> fixed point conversions * 31 30 29 28 24 23 22 21 20 19 18 16 15 10 9 5 4 0 * +----+---+---+-----------+------+---+-------+--------+-------+------+------+ * | sf | 0 | S | 1 1 1 1 0 | type | 0 | rmode | opcode | scale | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void handle_fmov(DisasContext *s, int rd, int rn, int type, bool itof) } } -/* C3.6.30 Floating point <-> integer conversions +/* Floating point <-> integer conversions * 31 30 29 28 24 23 22 21 20 19 18 16 15 10 9 5 4 0 * +----+---+---+-----------+------+---+-------+-----+-------------+----+----+ * | sf | 0 | S | 1 1 1 1 0 | type | 1 | rmode | opc | 0 0 0 0 0 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void do_ext64(DisasContext *s, TCGv_i64 tcg_left, TCGv_i64 tcg_right, tcg_temp_free_i64(tcg_tmp); } -/* C3.6.1 EXT +/* EXT * 31 30 29 24 23 22 21 20 16 15 14 11 10 9 5 4 0 * +---+---+-------------+-----+---+------+---+------+---+------+------+ * | 0 | Q | 1 0 1 1 1 0 | op2 | 0 | Rm | 0 | imm4 | 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_simd_ext(DisasContext *s, uint32_t insn) tcg_temp_free_i64(tcg_resh); } -/* C3.6.2 TBL/TBX +/* TBL/TBX * 31 30 29 24 23 22 21 20 16 15 14 13 12 11 10 9 5 4 0 * +---+---+-------------+-----+---+------+---+-----+----+-----+------+------+ * | 0 | Q | 0 0 1 1 1 0 | op2 | 0 | Rm | 0 | len | op | 0 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_simd_tb(DisasContext *s, uint32_t insn) tcg_temp_free_i64(tcg_resh); } -/* C3.6.3 ZIP/UZP/TRN +/* ZIP/UZP/TRN * 31 30 29 24 23 22 21 20 16 15 14 12 11 10 9 5 4 0 * +---+---+-------------+------+---+------+---+------------------+------+ * | 0 | Q | 0 0 1 1 1 0 | size | 0 | Rm | 0 | opc | 1 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void do_minmaxop(DisasContext *s, TCGv_i32 tcg_elt1, TCGv_i32 tcg_elt2, } } -/* C3.6.4 AdvSIMD across lanes +/* AdvSIMD across lanes * 31 30 29 28 24 23 22 21 17 16 12 11 10 9 5 4 0 * +---+---+---+-----------+------+-----------+--------+-----+------+------+ * | 0 | Q | U | 0 1 1 1 0 | size | 1 1 0 0 0 | opcode | 1 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_simd_across_lanes(DisasContext *s, uint32_t insn) tcg_temp_free_i64(tcg_res); } -/* C6.3.31 DUP (Element, Vector) +/* DUP (Element, Vector) * * 31 30 29 21 20 16 15 10 9 5 4 0 * +---+---+-------------------+--------+-------------+------+------+ @@ -XXX,XX +XXX,XX @@ static void handle_simd_dupe(DisasContext *s, int is_q, int rd, int rn, tcg_temp_free_i64(tmp); } -/* C6.3.31 DUP (element, scalar) +/* DUP (element, scalar) * 31 21 20 16 15 10 9 5 4 0 * +-----------------------+--------+-------------+------+------+ * | 0 1 0 1 1 1 1 0 0 0 0 | imm5 | 0 0 0 0 0 1 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void handle_simd_dupes(DisasContext *s, int rd, int rn, tcg_temp_free_i64(tmp); } -/* C6.3.32 DUP (General) +/* DUP (General) * * 31 30 29 21 20 16 15 10 9 5 4 0 * +---+---+-------------------+--------+-------------+------+------+ @@ -XXX,XX +XXX,XX @@ static void handle_simd_dupg(DisasContext *s, int is_q, int rd, int rn, } } -/* C6.3.150 INS (Element) +/* INS (Element) * * 31 21 20 16 15 14 11 10 9 5 4 0 * +-----------------------+--------+------------+---+------+------+ @@ -XXX,XX +XXX,XX @@ static void handle_simd_inse(DisasContext *s, int rd, int rn, } -/* C6.3.151 INS (General) +/* INS (General) * * 31 21 20 16 15 10 9 5 4 0 * +-----------------------+--------+-------------+------+------+ @@ -XXX,XX +XXX,XX @@ static void handle_simd_insg(DisasContext *s, int rd, int rn, int imm5) } /* - * C6.3.321 UMOV (General) - * C6.3.237 SMOV (General) + * UMOV (General) + * SMOV (General) * * 31 30 29 21 20 16 15 12 10 9 5 4 0 * +---+---+-------------------+--------+-------------+------+------+ @@ -XXX,XX +XXX,XX @@ static void handle_simd_umov_smov(DisasContext *s, int is_q, int is_signed, } } -/* C3.6.5 AdvSIMD copy +/* AdvSIMD copy * 31 30 29 28 21 20 16 15 14 11 10 9 5 4 0 * +---+---+----+-----------------+------+---+------+---+------+------+ * | 0 | Q | op | 0 1 1 1 0 0 0 0 | imm5 | 0 | imm4 | 1 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_simd_copy(DisasContext *s, uint32_t insn) } } -/* C3.6.6 AdvSIMD modified immediate +/* AdvSIMD modified immediate * 31 30 29 28 19 18 16 15 12 11 10 9 5 4 0 * +---+---+----+---------------------+-----+-------+----+---+-------+------+ * | 0 | Q | op | 0 1 1 1 1 0 0 0 0 0 | abc | cmode | o2 | 1 | defgh | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_simd_mod_imm(DisasContext *s, uint32_t insn) tcg_temp_free_i64(tcg_imm); } -/* C3.6.7 AdvSIMD scalar copy +/* AdvSIMD scalar copy * 31 30 29 28 21 20 16 15 14 11 10 9 5 4 0 * +-----+----+-----------------+------+---+------+---+------+------+ * | 0 1 | op | 1 1 1 1 0 0 0 0 | imm5 | 0 | imm4 | 1 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_simd_scalar_copy(DisasContext *s, uint32_t insn) handle_simd_dupes(s, rd, rn, imm5); } -/* C3.6.8 AdvSIMD scalar pairwise +/* AdvSIMD scalar pairwise * 31 30 29 28 24 23 22 21 17 16 12 11 10 9 5 4 0 * +-----+---+-----------+------+-----------+--------+-----+------+------+ * | 0 1 | U | 1 1 1 1 0 | size | 1 1 0 0 0 | opcode | 1 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void handle_simd_shift_fpint_conv(DisasContext *s, bool is_scalar, tcg_temp_free_i32(tcg_rmode); } -/* C3.6.9 AdvSIMD scalar shift by immediate +/* AdvSIMD scalar shift by immediate * 31 30 29 28 23 22 19 18 16 15 11 10 9 5 4 0 * +-----+---+-------------+------+------+--------+---+------+------+ * | 0 1 | U | 1 1 1 1 1 0 | immh | immb | opcode | 1 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_simd_scalar_shift_imm(DisasContext *s, uint32_t insn) } } -/* C3.6.10 AdvSIMD scalar three different +/* AdvSIMD scalar three different * 31 30 29 28 24 23 22 21 20 16 15 12 11 10 9 5 4 0 * +-----+---+-----------+------+---+------+--------+-----+------+------+ * | 0 1 | U | 1 1 1 1 0 | size | 1 | Rm | opcode | 0 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void handle_3same_float(DisasContext *s, int size, int elements, } } -/* C3.6.11 AdvSIMD scalar three same +/* AdvSIMD scalar three same * 31 30 29 28 24 23 22 21 20 16 15 11 10 9 5 4 0 * +-----+---+-----------+------+---+------+--------+---+------+------+ * | 0 1 | U | 1 1 1 1 0 | size | 1 | Rm | opcode | 1 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void handle_2misc_satacc(DisasContext *s, bool is_scalar, bool is_u, } } -/* C3.6.12 AdvSIMD scalar two reg misc +/* AdvSIMD scalar two reg misc * 31 30 29 28 24 23 22 21 17 16 12 11 10 9 5 4 0 * +-----+---+-----------+------+-----------+--------+-----+------+------+ * | 0 1 | U | 1 1 1 1 0 | size | 1 0 0 0 0 | opcode | 1 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void handle_vec_simd_shrn(DisasContext *s, bool is_q, } -/* C3.6.14 AdvSIMD shift by immediate +/* AdvSIMD shift by immediate * 31 30 29 28 23 22 19 18 16 15 11 10 9 5 4 0 * +---+---+---+-------------+------+------+--------+---+------+------+ * | 0 | Q | U | 0 1 1 1 1 0 | immh | immb | opcode | 1 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void handle_pmull_64(DisasContext *s, int is_q, int rd, int rn, int rm) tcg_temp_free_i64(tcg_res); } -/* C3.6.15 AdvSIMD three different +/* AdvSIMD three different * 31 30 29 28 24 23 22 21 20 16 15 12 11 10 9 5 4 0 * +---+---+---+-----------+------+---+------+--------+-----+------+------+ * | 0 | Q | U | 0 1 1 1 0 | size | 1 | Rm | opcode | 0 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_simd_3same_int(DisasContext *s, uint32_t insn) } } -/* C3.6.16 AdvSIMD three same +/* AdvSIMD three same * 31 30 29 28 24 23 22 21 20 16 15 11 10 9 5 4 0 * +---+---+---+-----------+------+---+------+--------+---+------+------+ * | 0 | Q | U | 0 1 1 1 0 | size | 1 | Rm | opcode | 1 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void handle_shll(DisasContext *s, bool is_q, int size, int rn, int rd) } } -/* C3.6.17 AdvSIMD two reg misc +/* AdvSIMD two reg misc * 31 30 29 28 24 23 22 21 17 16 12 11 10 9 5 4 0 * +---+---+---+-----------+------+-----------+--------+-----+------+------+ * | 0 | Q | U | 0 1 1 1 0 | size | 1 0 0 0 0 | opcode | 1 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc(DisasContext *s, uint32_t insn) } } -/* C3.6.13 AdvSIMD scalar x indexed element +/* AdvSIMD scalar x indexed element * 31 30 29 28 24 23 22 21 20 19 16 15 12 11 10 9 5 4 0 * +-----+---+-----------+------+---+---+------+-----+---+---+------+------+ * | 0 1 | U | 1 1 1 1 1 | size | L | M | Rm | opc | H | 0 | Rn | Rd | * +-----+---+-----------+------+---+---+------+-----+---+---+------+------+ - * C3.6.18 AdvSIMD vector x indexed element + * AdvSIMD vector x indexed element * 31 30 29 28 24 23 22 21 20 19 16 15 12 11 10 9 5 4 0 * +---+---+---+-----------+------+---+---+------+-----+---+---+------+------+ * | 0 | Q | U | 0 1 1 1 1 | size | L | M | Rm | opc | H | 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_simd_indexed(DisasContext *s, uint32_t insn) } } -/* C3.6.19 Crypto AES +/* Crypto AES * 31 24 23 22 21 17 16 12 11 10 9 5 4 0 * +-----------------+------+-----------+--------+-----+------+------+ * | 0 1 0 0 1 1 1 0 | size | 1 0 1 0 0 | opcode | 1 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_crypto_aes(DisasContext *s, uint32_t insn) tcg_temp_free_i32(tcg_decrypt); } -/* C3.6.20 Crypto three-reg SHA +/* Crypto three-reg SHA * 31 24 23 22 21 20 16 15 14 12 11 10 9 5 4 0 * +-----------------+------+---+------+---+--------+-----+------+------+ * | 0 1 0 1 1 1 1 0 | size | 0 | Rm | 0 | opcode | 0 0 | Rn | Rd | @@ -XXX,XX +XXX,XX @@ static void disas_crypto_three_reg_sha(DisasContext *s, uint32_t insn) tcg_temp_free_i32(tcg_rm_regno); } -/* C3.6.21 Crypto two-reg SHA +/* Crypto two-reg SHA * 31 24 23 22 21 17 16 12 11 10 9 5 4 0 * +-----------------+------+-----------+--------+-----+------+------+ * | 0 1 0 1 1 1 1 0 | size | 1 0 1 0 0 | opcode | 1 0 | Rn | Rd | -- 2.7.4 Update the static_ops functions to use new-style mmio rather than the legacy old_mmio functions. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505580378-9044-2-git-send-email-peter.maydell@linaro.org --- hw/arm/palm.c | 30 ++++++++++-------------------- 1 file changed, 10 insertions(+), 20 deletions(-) diff --git a/hw/arm/palm.c b/hw/arm/palm.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/palm.c +++ b/hw/arm/palm.c @@ -XXX,XX +XXX,XX @@ #include "exec/address-spaces.h" #include "cpu.h" -static uint32_t static_readb(void *opaque, hwaddr offset) +static uint64_t static_read(void *opaque, hwaddr offset, unsigned size) { - uint32_t *val = (uint32_t *) opaque; - return *val >> ((offset & 3) << 3); -} + uint32_t *val = (uint32_t *)opaque; + uint32_t sizemask = 7 >> size; -static uint32_t static_readh(void *opaque, hwaddr offset) -{ - uint32_t *val = (uint32_t *) opaque; - return *val >> ((offset & 1) << 3); -} - -static uint32_t static_readw(void *opaque, hwaddr offset) -{ - uint32_t *val = (uint32_t *) opaque; - return *val >> ((offset & 0) << 3); + return *val >> ((offset & sizemask) << 3); } -static void static_write(void *opaque, hwaddr offset, - uint32_t value) +static void static_write(void *opaque, hwaddr offset, uint64_t value, + unsigned size) { #ifdef SPY printf("%s: value %08lx written at " PA_FMT "\n", @@ -XXX,XX +XXX,XX @@ static void static_write(void *opaque, hwaddr offset, } static const MemoryRegionOps static_ops = { - .old_mmio = { - .read = { static_readb, static_readh, static_readw, }, - .write = { static_write, static_write, static_write, }, - }, + .read = static_read, + .write = static_write, + .valid.min_access_size = 1, + .valid.max_access_size = 4, .endianness = DEVICE_NATIVE_ENDIAN, }; -- 2.7.4 Drop the use of old_mmio in the omap2_gpio memory ops. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505580378-9044-3-git-send-email-peter.maydell@linaro.org --- hw/gpio/omap_gpio.c | 26 ++++++++++++-------------- 1 file changed, 12 insertions(+), 14 deletions(-) diff --git a/hw/gpio/omap_gpio.c b/hw/gpio/omap_gpio.c index XXXXXXX..XXXXXXX 100644 --- a/hw/gpio/omap_gpio.c +++ b/hw/gpio/omap_gpio.c @@ -XXX,XX +XXX,XX @@ static void omap2_gpio_module_write(void *opaque, hwaddr addr, } } -static uint32_t omap2_gpio_module_readp(void *opaque, hwaddr addr) +static uint64_t omap2_gpio_module_readp(void *opaque, hwaddr addr, + unsigned size) { return omap2_gpio_module_read(opaque, addr & ~3) >> ((addr & 3) << 3); } static void omap2_gpio_module_writep(void *opaque, hwaddr addr, - uint32_t value) + uint64_t value, unsigned size) { uint32_t cur = 0; uint32_t mask = 0xffff; + if (size == 4) { + omap2_gpio_module_write(opaque, addr, value); + return; + } + switch (addr & ~3) { case 0x00: /* GPIO_REVISION */ case 0x14: /* GPIO_SYSSTATUS */ @@ -XXX,XX +XXX,XX @@ static void omap2_gpio_module_writep(void *opaque, hwaddr addr, } static const MemoryRegionOps omap2_gpio_module_ops = { - .old_mmio = { - .read = { - omap2_gpio_module_readp, - omap2_gpio_module_readp, - omap2_gpio_module_read, - }, - .write = { - omap2_gpio_module_writep, - omap2_gpio_module_writep, - omap2_gpio_module_write, - }, - }, + .read = omap2_gpio_module_readp, + .write = omap2_gpio_module_writep, + .valid.min_access_size = 1, + .valid.max_access_size = 4, .endianness = DEVICE_NATIVE_ENDIAN, }; -- 2.7.4 Don't use the old_mmio in the memory region ops struct. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505580378-9044-4-git-send-email-peter.maydell@linaro.org --- hw/timer/omap_synctimer.c | 35 +++++++++++++++++++++-------------- 1 file changed, 21 insertions(+), 14 deletions(-) diff --git a/hw/timer/omap_synctimer.c b/hw/timer/omap_synctimer.c index XXXXXXX..XXXXXXX 100644 --- a/hw/timer/omap_synctimer.c +++ b/hw/timer/omap_synctimer.c @@ -XXX,XX +XXX,XX @@ static uint32_t omap_synctimer_readh(void *opaque, hwaddr addr) } } -static void omap_synctimer_write(void *opaque, hwaddr addr, - uint32_t value) +static uint64_t omap_synctimer_readfn(void *opaque, hwaddr addr, + unsigned size) +{ + switch (size) { + case 1: + return omap_badwidth_read32(opaque, addr); + case 2: + return omap_synctimer_readh(opaque, addr); + case 4: + return omap_synctimer_readw(opaque, addr); + default: + g_assert_not_reached(); + } +} + +static void omap_synctimer_writefn(void *opaque, hwaddr addr, + uint64_t value, unsigned size) { OMAP_BAD_REG(addr); } static const MemoryRegionOps omap_synctimer_ops = { - .old_mmio = { - .read = { - omap_badwidth_read32, - omap_synctimer_readh, - omap_synctimer_readw, - }, - .write = { - omap_badwidth_write32, - omap_synctimer_write, - omap_synctimer_write, - }, - }, + .read = omap_synctimer_readfn, + .write = omap_synctimer_writefn, + .valid.min_access_size = 1, + .valid.max_access_size = 4, .endianness = DEVICE_NATIVE_ENDIAN, }; -- 2.7.4 Don't use the old_mmio struct in memory region ops. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505580378-9044-5-git-send-email-peter.maydell@linaro.org --- hw/timer/omap_gptimer.c | 49 +++++++++++++++++++++++++++++++++++++------------ 1 file changed, 37 insertions(+), 12 deletions(-) diff --git a/hw/timer/omap_gptimer.c b/hw/timer/omap_gptimer.c index XXXXXXX..XXXXXXX 100644 --- a/hw/timer/omap_gptimer.c +++ b/hw/timer/omap_gptimer.c @@ -XXX,XX +XXX,XX @@ static void omap_gp_timer_writeh(void *opaque, hwaddr addr, s->writeh = (uint16_t) value; } +static uint64_t omap_gp_timer_readfn(void *opaque, hwaddr addr, + unsigned size) +{ + switch (size) { + case 1: + return omap_badwidth_read32(opaque, addr); + case 2: + return omap_gp_timer_readh(opaque, addr); + case 4: + return omap_gp_timer_readw(opaque, addr); + default: + g_assert_not_reached(); + } +} + +static void omap_gp_timer_writefn(void *opaque, hwaddr addr, + uint64_t value, unsigned size) +{ + switch (size) { + case 1: + omap_badwidth_write32(opaque, addr, value); + break; + case 2: + omap_gp_timer_writeh(opaque, addr, value); + break; + case 4: + omap_gp_timer_write(opaque, addr, value); + break; + default: + g_assert_not_reached(); + } +} + static const MemoryRegionOps omap_gp_timer_ops = { - .old_mmio = { - .read = { - omap_badwidth_read32, - omap_gp_timer_readh, - omap_gp_timer_readw, - }, - .write = { - omap_badwidth_write32, - omap_gp_timer_writeh, - omap_gp_timer_write, - }, - }, + .read = omap_gp_timer_readfn, + .write = omap_gp_timer_writefn, + .valid.min_access_size = 1, + .valid.max_access_size = 4, .endianness = DEVICE_NATIVE_ENDIAN, }; -- 2.7.4 Don't use old_mmio in the memory region ops struct. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505580378-9044-6-git-send-email-peter.maydell@linaro.org --- hw/i2c/omap_i2c.c | 44 ++++++++++++++++++++++++++++++++------------ 1 file changed, 32 insertions(+), 12 deletions(-) diff --git a/hw/i2c/omap_i2c.c b/hw/i2c/omap_i2c.c index XXXXXXX..XXXXXXX 100644 --- a/hw/i2c/omap_i2c.c +++ b/hw/i2c/omap_i2c.c @@ -XXX,XX +XXX,XX @@ static void omap_i2c_writeb(void *opaque, hwaddr addr, } } +static uint64_t omap_i2c_readfn(void *opaque, hwaddr addr, + unsigned size) +{ + switch (size) { + case 2: + return omap_i2c_read(opaque, addr); + default: + return omap_badwidth_read16(opaque, addr); + } +} + +static void omap_i2c_writefn(void *opaque, hwaddr addr, + uint64_t value, unsigned size) +{ + switch (size) { + case 1: + /* Only the last fifo write can be 8 bit. */ + omap_i2c_writeb(opaque, addr, value); + break; + case 2: + omap_i2c_write(opaque, addr, value); + break; + default: + omap_badwidth_write16(opaque, addr, value); + break; + } +} + static const MemoryRegionOps omap_i2c_ops = { - .old_mmio = { - .read = { - omap_badwidth_read16, - omap_i2c_read, - omap_badwidth_read16, - }, - .write = { - omap_i2c_writeb, /* Only the last fifo write can be 8 bit. */ - omap_i2c_write, - omap_badwidth_write16, - }, - }, + .read = omap_i2c_readfn, + .write = omap_i2c_writefn, + .valid.min_access_size = 1, + .valid.max_access_size = 4, .endianness = DEVICE_NATIVE_ENDIAN, }; -- 2.7.4 Don't use old_mmio in the memory region ops struct. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 1505580378-9044-7-git-send-email-peter.maydell@linaro.org --- hw/arm/omap2.c | 49 +++++++++++++++++++++++++++++++++++++------------ 1 file changed, 37 insertions(+), 12 deletions(-) diff --git a/hw/arm/omap2.c b/hw/arm/omap2.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/omap2.c +++ b/hw/arm/omap2.c @@ -XXX,XX +XXX,XX @@ static void omap_sysctl_write(void *opaque, hwaddr addr, } } +static uint64_t omap_sysctl_readfn(void *opaque, hwaddr addr, + unsigned size) +{ + switch (size) { + case 1: + return omap_sysctl_read8(opaque, addr); + case 2: + return omap_badwidth_read32(opaque, addr); /* TODO */ + case 4: + return omap_sysctl_read(opaque, addr); + default: + g_assert_not_reached(); + } +} + +static void omap_sysctl_writefn(void *opaque, hwaddr addr, + uint64_t value, unsigned size) +{ + switch (size) { + case 1: + omap_sysctl_write8(opaque, addr, value); + break; + case 2: + omap_badwidth_write32(opaque, addr, value); /* TODO */ + break; + case 4: + omap_sysctl_write(opaque, addr, value); + break; + default: + g_assert_not_reached(); + } +} + static const MemoryRegionOps omap_sysctl_ops = { - .old_mmio = { - .read = { - omap_sysctl_read8, - omap_badwidth_read32, /* TODO */ - omap_sysctl_read, - }, - .write = { - omap_sysctl_write8, - omap_badwidth_write32, /* TODO */ - omap_sysctl_write, - }, - }, + .read = omap_sysctl_readfn, + .write = omap_sysctl_writefn, + .valid.min_access_size = 1, + .valid.max_access_size = 4, .endianness = DEVICE_NATIVE_ENDIAN, }; -- 2.7.4 From: Subbaraya Sundeep <sundeep.lkml@gmail.com> Modelled System Timer in Microsemi's Smartfusion2 Soc. Timer has two 32bit down counters and two interrupts. Signed-off-by: Subbaraya Sundeep <sundeep.lkml@gmail.com> Reviewed-by: Alistair Francis <alistair.francis@xilinx.com> Acked-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20170920201737.25723-2-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/timer/Makefile.objs | 1 + include/hw/timer/mss-timer.h | 64 ++++++++++ hw/timer/mss-timer.c | 289 +++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 354 insertions(+) create mode 100644 include/hw/timer/mss-timer.h create mode 100644 hw/timer/mss-timer.c diff --git a/hw/timer/Makefile.objs b/hw/timer/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/hw/timer/Makefile.objs +++ b/hw/timer/Makefile.objs @@ -XXX,XX +XXX,XX @@ common-obj-$(CONFIG_ASPEED_SOC) += aspeed_timer.o common-obj-$(CONFIG_SUN4V_RTC) += sun4v-rtc.o common-obj-$(CONFIG_CMSDK_APB_TIMER) += cmsdk-apb-timer.o +common-obj-$(CONFIG_MSF2) += mss-timer.o diff --git a/include/hw/timer/mss-timer.h b/include/hw/timer/mss-timer.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/timer/mss-timer.h @@ -XXX,XX +XXX,XX @@ +/* + * Microsemi SmartFusion2 Timer. + * + * Copyright (c) 2017 Subbaraya Sundeep <sundeep.lkml@gmail.com> + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + */ + +#ifndef HW_MSS_TIMER_H +#define HW_MSS_TIMER_H + +#include "hw/sysbus.h" +#include "hw/ptimer.h" + +#define TYPE_MSS_TIMER "mss-timer" +#define MSS_TIMER(obj) OBJECT_CHECK(MSSTimerState, \ + (obj), TYPE_MSS_TIMER) + +/* + * There are two 32-bit down counting timers. + * Timers 1 and 2 can be concatenated into a single 64-bit Timer + * that operates either in Periodic mode or in One-shot mode. + * Writing 1 to the TIM64_MODE register bit 0 sets the Timers in 64-bit mode. + * In 64-bit mode, writing to the 32-bit registers has no effect. + * Similarly, in 32-bit mode, writing to the 64-bit mode registers + * has no effect. Only two 32-bit timers are supported currently. + */ +#define NUM_TIMERS 2 + +#define R_TIM1_MAX 6 + +struct Msf2Timer { + QEMUBH *bh; + ptimer_state *ptimer; + + uint32_t regs[R_TIM1_MAX]; + qemu_irq irq; +}; + +typedef struct MSSTimerState { + SysBusDevice parent_obj; + + MemoryRegion mmio; + uint32_t freq_hz; + struct Msf2Timer timers[NUM_TIMERS]; +} MSSTimerState; + +#endif /* HW_MSS_TIMER_H */ diff --git a/hw/timer/mss-timer.c b/hw/timer/mss-timer.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/timer/mss-timer.c @@ -XXX,XX +XXX,XX @@ +/* + * Block model of System timer present in + * Microsemi's SmartFusion2 and SmartFusion SoCs. + * + * Copyright (c) 2017 Subbaraya Sundeep <sundeep.lkml@gmail.com>. + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + */ + +#include "qemu/osdep.h" +#include "qemu/main-loop.h" +#include "qemu/log.h" +#include "hw/timer/mss-timer.h" + +#ifndef MSS_TIMER_ERR_DEBUG +#define MSS_TIMER_ERR_DEBUG 0 +#endif + +#define DB_PRINT_L(lvl, fmt, args...) do { \ + if (MSS_TIMER_ERR_DEBUG >= lvl) { \ + qemu_log("%s: " fmt "\n", __func__, ## args); \ + } \ +} while (0); + +#define DB_PRINT(fmt, args...) DB_PRINT_L(1, fmt, ## args) + +#define R_TIM_VAL 0 +#define R_TIM_LOADVAL 1 +#define R_TIM_BGLOADVAL 2 +#define R_TIM_CTRL 3 +#define R_TIM_RIS 4 +#define R_TIM_MIS 5 + +#define TIMER_CTRL_ENBL (1 << 0) +#define TIMER_CTRL_ONESHOT (1 << 1) +#define TIMER_CTRL_INTR (1 << 2) +#define TIMER_RIS_ACK (1 << 0) +#define TIMER_RST_CLR (1 << 6) +#define TIMER_MODE (1 << 0) + +static void timer_update_irq(struct Msf2Timer *st) +{ + bool isr, ier; + + isr = !!(st->regs[R_TIM_RIS] & TIMER_RIS_ACK); + ier = !!(st->regs[R_TIM_CTRL] & TIMER_CTRL_INTR); + qemu_set_irq(st->irq, (ier && isr)); +} + +static void timer_update(struct Msf2Timer *st) +{ + uint64_t count; + + if (!(st->regs[R_TIM_CTRL] & TIMER_CTRL_ENBL)) { + ptimer_stop(st->ptimer); + return; + } + + count = st->regs[R_TIM_LOADVAL]; + ptimer_set_limit(st->ptimer, count, 1); + ptimer_run(st->ptimer, 1); +} + +static uint64_t +timer_read(void *opaque, hwaddr offset, unsigned int size) +{ + MSSTimerState *t = opaque; + hwaddr addr; + struct Msf2Timer *st; + uint32_t ret = 0; + int timer = 0; + int isr; + int ier; + + addr = offset >> 2; + /* + * Two independent timers has same base address. + * Based on address passed figure out which timer is being used. + */ + if ((addr >= R_TIM1_MAX) && (addr < NUM_TIMERS * R_TIM1_MAX)) { + timer = 1; + addr -= R_TIM1_MAX; + } + + st = &t->timers[timer]; + + switch (addr) { + case R_TIM_VAL: + ret = ptimer_get_count(st->ptimer); + break; + + case R_TIM_MIS: + isr = !!(st->regs[R_TIM_RIS] & TIMER_RIS_ACK); + ier = !!(st->regs[R_TIM_CTRL] & TIMER_CTRL_INTR); + ret = ier & isr; + break; + + default: + if (addr < R_TIM1_MAX) { + ret = st->regs[addr]; + } else { + qemu_log_mask(LOG_GUEST_ERROR, + TYPE_MSS_TIMER": 64-bit mode not supported\n"); + return ret; + } + break; + } + + DB_PRINT("timer=%d 0x%" HWADDR_PRIx "=0x%" PRIx32, timer, offset, + ret); + return ret; +} + +static void +timer_write(void *opaque, hwaddr offset, + uint64_t val64, unsigned int size) +{ + MSSTimerState *t = opaque; + hwaddr addr; + struct Msf2Timer *st; + int timer = 0; + uint32_t value = val64; + + addr = offset >> 2; + /* + * Two independent timers has same base address. + * Based on addr passed figure out which timer is being used. + */ + if ((addr >= R_TIM1_MAX) && (addr < NUM_TIMERS * R_TIM1_MAX)) { + timer = 1; + addr -= R_TIM1_MAX; + } + + st = &t->timers[timer]; + + DB_PRINT("addr=0x%" HWADDR_PRIx " val=0x%" PRIx32 " (timer=%d)", offset, + value, timer); + + switch (addr) { + case R_TIM_CTRL: + st->regs[R_TIM_CTRL] = value; + timer_update(st); + break; + + case R_TIM_RIS: + if (value & TIMER_RIS_ACK) { + st->regs[R_TIM_RIS] &= ~TIMER_RIS_ACK; + } + break; + + case R_TIM_LOADVAL: + st->regs[R_TIM_LOADVAL] = value; + if (st->regs[R_TIM_CTRL] & TIMER_CTRL_ENBL) { + timer_update(st); + } + break; + + case R_TIM_BGLOADVAL: + st->regs[R_TIM_BGLOADVAL] = value; + st->regs[R_TIM_LOADVAL] = value; + break; + + case R_TIM_VAL: + case R_TIM_MIS: + break; + + default: + if (addr < R_TIM1_MAX) { + st->regs[addr] = value; + } else { + qemu_log_mask(LOG_GUEST_ERROR, + TYPE_MSS_TIMER": 64-bit mode not supported\n"); + return; + } + break; + } + timer_update_irq(st); +} + +static const MemoryRegionOps timer_ops = { + .read = timer_read, + .write = timer_write, + .endianness = DEVICE_NATIVE_ENDIAN, + .valid = { + .min_access_size = 1, + .max_access_size = 4 + } +}; + +static void timer_hit(void *opaque) +{ + struct Msf2Timer *st = opaque; + + st->regs[R_TIM_RIS] |= TIMER_RIS_ACK; + + if (!(st->regs[R_TIM_CTRL] & TIMER_CTRL_ONESHOT)) { + timer_update(st); + } + timer_update_irq(st); +} + +static void mss_timer_init(Object *obj) +{ + MSSTimerState *t = MSS_TIMER(obj); + int i; + + /* Init all the ptimers. */ + for (i = 0; i < NUM_TIMERS; i++) { + struct Msf2Timer *st = &t->timers[i]; + + st->bh = qemu_bh_new(timer_hit, st); + st->ptimer = ptimer_init(st->bh, PTIMER_POLICY_DEFAULT); + ptimer_set_freq(st->ptimer, t->freq_hz); + sysbus_init_irq(SYS_BUS_DEVICE(obj), &st->irq); + } + + memory_region_init_io(&t->mmio, OBJECT(t), &timer_ops, t, TYPE_MSS_TIMER, + NUM_TIMERS * R_TIM1_MAX * 4); + sysbus_init_mmio(SYS_BUS_DEVICE(obj), &t->mmio); +} + +static const VMStateDescription vmstate_timers = { + .name = "mss-timer-block", + .version_id = 1, + .minimum_version_id = 1, + .fields = (VMStateField[]) { + VMSTATE_PTIMER(ptimer, struct Msf2Timer), + VMSTATE_UINT32_ARRAY(regs, struct Msf2Timer, R_TIM1_MAX), + VMSTATE_END_OF_LIST() + } +}; + +static const VMStateDescription vmstate_mss_timer = { + .name = TYPE_MSS_TIMER, + .version_id = 1, + .minimum_version_id = 1, + .fields = (VMStateField[]) { + VMSTATE_UINT32(freq_hz, MSSTimerState), + VMSTATE_STRUCT_ARRAY(timers, MSSTimerState, NUM_TIMERS, 0, + vmstate_timers, struct Msf2Timer), + VMSTATE_END_OF_LIST() + } +}; + +static Property mss_timer_properties[] = { + /* Libero GUI shows 100Mhz as default for clocks */ + DEFINE_PROP_UINT32("clock-frequency", MSSTimerState, freq_hz, + 100 * 1000000), + DEFINE_PROP_END_OF_LIST(), +}; + +static void mss_timer_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + + dc->props = mss_timer_properties; + dc->vmsd = &vmstate_mss_timer; +} + +static const TypeInfo mss_timer_info = { + .name = TYPE_MSS_TIMER, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(MSSTimerState), + .instance_init = mss_timer_init, + .class_init = mss_timer_class_init, +}; + +static void mss_timer_register_types(void) +{ + type_register_static(&mss_timer_info); +} + +type_init(mss_timer_register_types) -- 2.7.4 From: Subbaraya Sundeep <sundeep.lkml@gmail.com> Added Sytem register block of Smartfusion2. This block has PLL registers which are accessed by guest. Signed-off-by: Subbaraya Sundeep <sundeep.lkml@gmail.com> Reviewed-by: Alistair Francis <alistair.francis@xilinx.com> Acked-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20170920201737.25723-3-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/misc/Makefile.objs | 1 + include/hw/misc/msf2-sysreg.h | 77 ++++++++++++++++++++ hw/misc/msf2-sysreg.c | 160 ++++++++++++++++++++++++++++++++++++++++++ hw/misc/trace-events | 5 ++ 4 files changed, 243 insertions(+) create mode 100644 include/hw/misc/msf2-sysreg.h create mode 100644 hw/misc/msf2-sysreg.c diff --git a/hw/misc/Makefile.objs b/hw/misc/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/Makefile.objs +++ b/hw/misc/Makefile.objs @@ -XXX,XX +XXX,XX @@ obj-$(CONFIG_HYPERV_TESTDEV) += hyperv_testdev.o obj-$(CONFIG_AUX) += auxbus.o obj-$(CONFIG_ASPEED_SOC) += aspeed_scu.o aspeed_sdmc.o obj-y += mmio_interface.o +obj-$(CONFIG_MSF2) += msf2-sysreg.o diff --git a/include/hw/misc/msf2-sysreg.h b/include/hw/misc/msf2-sysreg.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/misc/msf2-sysreg.h @@ -XXX,XX +XXX,XX @@ +/* + * Microsemi SmartFusion2 SYSREG + * + * Copyright (c) 2017 Subbaraya Sundeep <sundeep.lkml@gmail.com> + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + */ + +#ifndef HW_MSF2_SYSREG_H +#define HW_MSF2_SYSREG_H + +#include "hw/sysbus.h" + +enum { + ESRAM_CR = 0x00 / 4, + ESRAM_MAX_LAT, + DDR_CR, + ENVM_CR, + ENVM_REMAP_BASE_CR, + ENVM_REMAP_FAB_CR, + CC_CR, + CC_REGION_CR, + CC_LOCK_BASE_ADDR_CR, + CC_FLUSH_INDX_CR, + DDRB_BUF_TIMER_CR, + DDRB_NB_ADDR_CR, + DDRB_NB_SIZE_CR, + DDRB_CR, + + SOFT_RESET_CR = 0x48 / 4, + M3_CR, + + GPIO_SYSRESET_SEL_CR = 0x58 / 4, + + MDDR_CR = 0x60 / 4, + + MSSDDR_PLL_STATUS_LOW_CR = 0x90 / 4, + MSSDDR_PLL_STATUS_HIGH_CR, + MSSDDR_FACC1_CR, + MSSDDR_FACC2_CR, + + MSSDDR_PLL_STATUS = 0x150 / 4, +}; + +#define MSF2_SYSREG_MMIO_SIZE 0x300 + +#define TYPE_MSF2_SYSREG "msf2-sysreg" +#define MSF2_SYSREG(obj) OBJECT_CHECK(MSF2SysregState, (obj), TYPE_MSF2_SYSREG) + +typedef struct MSF2SysregState { + SysBusDevice parent_obj; + + MemoryRegion iomem; + + uint8_t apb0div; + uint8_t apb1div; + + uint32_t regs[MSF2_SYSREG_MMIO_SIZE / 4]; +} MSF2SysregState; + +#endif /* HW_MSF2_SYSREG_H */ diff --git a/hw/misc/msf2-sysreg.c b/hw/misc/msf2-sysreg.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/misc/msf2-sysreg.c @@ -XXX,XX +XXX,XX @@ +/* + * System Register block model of Microsemi SmartFusion2. + * + * Copyright (c) 2017 Subbaraya Sundeep <sundeep.lkml@gmail.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, see <http://www.gnu.org/licenses/>. + */ + +#include "qemu/osdep.h" +#include "qapi/error.h" +#include "qemu/log.h" +#include "hw/misc/msf2-sysreg.h" +#include "qemu/error-report.h" +#include "trace.h" + +static inline int msf2_divbits(uint32_t div) +{ + int r = ctz32(div); + + return (div < 8) ? r : r + 1; +} + +static void msf2_sysreg_reset(DeviceState *d) +{ + MSF2SysregState *s = MSF2_SYSREG(d); + + s->regs[MSSDDR_PLL_STATUS_LOW_CR] = 0x021A2358; + s->regs[MSSDDR_PLL_STATUS] = 0x3; + s->regs[MSSDDR_FACC1_CR] = msf2_divbits(s->apb0div) << 5 | + msf2_divbits(s->apb1div) << 2; +} + +static uint64_t msf2_sysreg_read(void *opaque, hwaddr offset, + unsigned size) +{ + MSF2SysregState *s = opaque; + uint32_t ret = 0; + + offset >>= 2; + if (offset < ARRAY_SIZE(s->regs)) { + ret = s->regs[offset]; + trace_msf2_sysreg_read(offset << 2, ret); + } else { + qemu_log_mask(LOG_GUEST_ERROR, + "%s: Bad offset 0x%08" HWADDR_PRIx "\n", __func__, + offset << 2); + } + + return ret; +} + +static void msf2_sysreg_write(void *opaque, hwaddr offset, + uint64_t val, unsigned size) +{ + MSF2SysregState *s = opaque; + uint32_t newval = val; + + offset >>= 2; + + switch (offset) { + case MSSDDR_PLL_STATUS: + trace_msf2_sysreg_write_pll_status(); + break; + + case ESRAM_CR: + case DDR_CR: + case ENVM_REMAP_BASE_CR: + if (newval != s->regs[offset]) { + qemu_log_mask(LOG_UNIMP, + TYPE_MSF2_SYSREG": remapping not supported\n"); + } + break; + + default: + if (offset < ARRAY_SIZE(s->regs)) { + trace_msf2_sysreg_write(offset << 2, newval, s->regs[offset]); + s->regs[offset] = newval; + } else { + qemu_log_mask(LOG_GUEST_ERROR, + "%s: Bad offset 0x%08" HWADDR_PRIx "\n", __func__, + offset << 2); + } + break; + } +} + +static const MemoryRegionOps sysreg_ops = { + .read = msf2_sysreg_read, + .write = msf2_sysreg_write, + .endianness = DEVICE_NATIVE_ENDIAN, +}; + +static void msf2_sysreg_init(Object *obj) +{ + MSF2SysregState *s = MSF2_SYSREG(obj); + + memory_region_init_io(&s->iomem, obj, &sysreg_ops, s, TYPE_MSF2_SYSREG, + MSF2_SYSREG_MMIO_SIZE); + sysbus_init_mmio(SYS_BUS_DEVICE(obj), &s->iomem); +} + +static const VMStateDescription vmstate_msf2_sysreg = { + .name = TYPE_MSF2_SYSREG, + .version_id = 1, + .minimum_version_id = 1, + .fields = (VMStateField[]) { + VMSTATE_UINT32_ARRAY(regs, MSF2SysregState, MSF2_SYSREG_MMIO_SIZE / 4), + VMSTATE_END_OF_LIST() + } +}; + +static Property msf2_sysreg_properties[] = { + /* default divisors in Libero GUI */ + DEFINE_PROP_UINT8("apb0divisor", MSF2SysregState, apb0div, 2), + DEFINE_PROP_UINT8("apb1divisor", MSF2SysregState, apb1div, 2), + DEFINE_PROP_END_OF_LIST(), +}; + +static void msf2_sysreg_realize(DeviceState *dev, Error **errp) +{ + MSF2SysregState *s = MSF2_SYSREG(dev); + + if ((s->apb0div > 32 || !is_power_of_2(s->apb0div)) + || (s->apb1div > 32 || !is_power_of_2(s->apb1div))) { + error_setg(errp, "Invalid apb divisor value"); + error_append_hint(errp, "apb divisor must be a power of 2" + " and maximum value is 32\n"); + } +} + +static void msf2_sysreg_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + + dc->vmsd = &vmstate_msf2_sysreg; + dc->reset = msf2_sysreg_reset; + dc->props = msf2_sysreg_properties; + dc->realize = msf2_sysreg_realize; +} + +static const TypeInfo msf2_sysreg_info = { + .name = TYPE_MSF2_SYSREG, + .parent = TYPE_SYS_BUS_DEVICE, + .class_init = msf2_sysreg_class_init, + .instance_size = sizeof(MSF2SysregState), + .instance_init = msf2_sysreg_init, +}; + +static void msf2_sysreg_register_types(void) +{ + type_register_static(&msf2_sysreg_info); +} + +type_init(msf2_sysreg_register_types) diff --git a/hw/misc/trace-events b/hw/misc/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/trace-events +++ b/hw/misc/trace-events @@ -XXX,XX +XXX,XX @@ mps2_scc_reset(void) "MPS2 SCC: reset" mps2_scc_leds(char led7, char led6, char led5, char led4, char led3, char led2, char led1, char led0) "MPS2 SCC LEDs: %c%c%c%c%c%c%c%c" mps2_scc_cfg_write(unsigned function, unsigned device, uint32_t value) "MPS2 SCC config write: function %d device %d data 0x%" PRIx32 mps2_scc_cfg_read(unsigned function, unsigned device, uint32_t value) "MPS2 SCC config read: function %d device %d data 0x%" PRIx32 + +# hw/misc/msf2-sysreg.c +msf2_sysreg_write(uint64_t offset, uint32_t val, uint32_t prev) "msf2-sysreg write: addr 0x%08" HWADDR_PRIx " data 0x%" PRIx32 " prev 0x%" PRIx32 +msf2_sysreg_read(uint64_t offset, uint32_t val) "msf2-sysreg read: addr 0x%08" HWADDR_PRIx " data 0x%08" PRIx32 +msf2_sysreg_write_pll_status(void) "Invalid write to read only PLL status register" -- 2.7.4 From: Subbaraya Sundeep <sundeep.lkml@gmail.com> Modelled Microsemi's Smartfusion2 SPI controller. Signed-off-by: Subbaraya Sundeep <sundeep.lkml@gmail.com> Reviewed-by: Alistair Francis <alistair.francis@xilinx.com> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20170920201737.25723-4-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/ssi/Makefile.objs | 1 + include/hw/ssi/mss-spi.h | 58 +++++++ hw/ssi/mss-spi.c | 404 +++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 463 insertions(+) create mode 100644 include/hw/ssi/mss-spi.h create mode 100644 hw/ssi/mss-spi.c diff --git a/hw/ssi/Makefile.objs b/hw/ssi/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/hw/ssi/Makefile.objs +++ b/hw/ssi/Makefile.objs @@ -XXX,XX +XXX,XX @@ common-obj-$(CONFIG_XILINX_SPI) += xilinx_spi.o common-obj-$(CONFIG_XILINX_SPIPS) += xilinx_spips.o common-obj-$(CONFIG_ASPEED_SOC) += aspeed_smc.o common-obj-$(CONFIG_STM32F2XX_SPI) += stm32f2xx_spi.o +common-obj-$(CONFIG_MSF2) += mss-spi.o obj-$(CONFIG_OMAP) += omap_spi.o obj-$(CONFIG_IMX) += imx_spi.o diff --git a/include/hw/ssi/mss-spi.h b/include/hw/ssi/mss-spi.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/ssi/mss-spi.h @@ -XXX,XX +XXX,XX @@ +/* + * Microsemi SmartFusion2 SPI + * + * Copyright (c) 2017 Subbaraya Sundeep <sundeep.lkml@gmail.com> + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + */ + +#ifndef HW_MSS_SPI_H +#define HW_MSS_SPI_H + +#include "hw/sysbus.h" +#include "hw/ssi/ssi.h" +#include "qemu/fifo32.h" + +#define TYPE_MSS_SPI "mss-spi" +#define MSS_SPI(obj) OBJECT_CHECK(MSSSpiState, (obj), TYPE_MSS_SPI) + +#define R_SPI_MAX 16 + +typedef struct MSSSpiState { + SysBusDevice parent_obj; + + MemoryRegion mmio; + + qemu_irq irq; + + qemu_irq cs_line; + + SSIBus *spi; + + Fifo32 rx_fifo; + Fifo32 tx_fifo; + + int fifo_depth; + uint32_t frame_count; + bool enabled; + + uint32_t regs[R_SPI_MAX]; +} MSSSpiState; + +#endif /* HW_MSS_SPI_H */ diff --git a/hw/ssi/mss-spi.c b/hw/ssi/mss-spi.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/ssi/mss-spi.c @@ -XXX,XX +XXX,XX @@ +/* + * Block model of SPI controller present in + * Microsemi's SmartFusion2 and SmartFusion SoCs. + * + * Copyright (C) 2017 Subbaraya Sundeep <sundeep.lkml@gmail.com> + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + */ + +#include "qemu/osdep.h" +#include "hw/ssi/mss-spi.h" +#include "qemu/log.h" + +#ifndef MSS_SPI_ERR_DEBUG +#define MSS_SPI_ERR_DEBUG 0 +#endif + +#define DB_PRINT_L(lvl, fmt, args...) do { \ + if (MSS_SPI_ERR_DEBUG >= lvl) { \ + qemu_log("%s: " fmt "\n", __func__, ## args); \ + } \ +} while (0); + +#define DB_PRINT(fmt, args...) DB_PRINT_L(1, fmt, ## args) + +#define FIFO_CAPACITY 32 + +#define R_SPI_CONTROL 0 +#define R_SPI_DFSIZE 1 +#define R_SPI_STATUS 2 +#define R_SPI_INTCLR 3 +#define R_SPI_RX 4 +#define R_SPI_TX 5 +#define R_SPI_CLKGEN 6 +#define R_SPI_SS 7 +#define R_SPI_MIS 8 +#define R_SPI_RIS 9 + +#define S_TXDONE (1 << 0) +#define S_RXRDY (1 << 1) +#define S_RXCHOVRF (1 << 2) +#define S_RXFIFOFUL (1 << 4) +#define S_RXFIFOFULNXT (1 << 5) +#define S_RXFIFOEMP (1 << 6) +#define S_RXFIFOEMPNXT (1 << 7) +#define S_TXFIFOFUL (1 << 8) +#define S_TXFIFOFULNXT (1 << 9) +#define S_TXFIFOEMP (1 << 10) +#define S_TXFIFOEMPNXT (1 << 11) +#define S_FRAMESTART (1 << 12) +#define S_SSEL (1 << 13) +#define S_ACTIVE (1 << 14) + +#define C_ENABLE (1 << 0) +#define C_MODE (1 << 1) +#define C_INTRXDATA (1 << 4) +#define C_INTTXDATA (1 << 5) +#define C_INTRXOVRFLO (1 << 6) +#define C_SPS (1 << 26) +#define C_BIGFIFO (1 << 29) +#define C_RESET (1 << 31) + +#define FRAMESZ_MASK 0x1F +#define FMCOUNT_MASK 0x00FFFF00 +#define FMCOUNT_SHIFT 8 + +static void txfifo_reset(MSSSpiState *s) +{ + fifo32_reset(&s->tx_fifo); + + s->regs[R_SPI_STATUS] &= ~S_TXFIFOFUL; + s->regs[R_SPI_STATUS] |= S_TXFIFOEMP; +} + +static void rxfifo_reset(MSSSpiState *s) +{ + fifo32_reset(&s->rx_fifo); + + s->regs[R_SPI_STATUS] &= ~S_RXFIFOFUL; + s->regs[R_SPI_STATUS] |= S_RXFIFOEMP; +} + +static void set_fifodepth(MSSSpiState *s) +{ + unsigned int size = s->regs[R_SPI_DFSIZE] & FRAMESZ_MASK; + + if (size <= 8) { + s->fifo_depth = 32; + } else if (size <= 16) { + s->fifo_depth = 16; + } else if (size <= 32) { + s->fifo_depth = 8; + } else { + s->fifo_depth = 4; + } +} + +static void update_mis(MSSSpiState *s) +{ + uint32_t reg = s->regs[R_SPI_CONTROL]; + uint32_t tmp; + + /* + * form the Control register interrupt enable bits + * same as RIS, MIS and Interrupt clear registers for simplicity + */ + tmp = ((reg & C_INTRXOVRFLO) >> 4) | ((reg & C_INTRXDATA) >> 3) | + ((reg & C_INTTXDATA) >> 5); + s->regs[R_SPI_MIS] |= tmp & s->regs[R_SPI_RIS]; +} + +static void spi_update_irq(MSSSpiState *s) +{ + int irq; + + update_mis(s); + irq = !!(s->regs[R_SPI_MIS]); + + qemu_set_irq(s->irq, irq); +} + +static void mss_spi_reset(DeviceState *d) +{ + MSSSpiState *s = MSS_SPI(d); + + memset(s->regs, 0, sizeof s->regs); + s->regs[R_SPI_CONTROL] = 0x80000102; + s->regs[R_SPI_DFSIZE] = 0x4; + s->regs[R_SPI_STATUS] = S_SSEL | S_TXFIFOEMP | S_RXFIFOEMP; + s->regs[R_SPI_CLKGEN] = 0x7; + s->regs[R_SPI_RIS] = 0x0; + + s->fifo_depth = 4; + s->frame_count = 1; + s->enabled = false; + + rxfifo_reset(s); + txfifo_reset(s); +} + +static uint64_t +spi_read(void *opaque, hwaddr addr, unsigned int size) +{ + MSSSpiState *s = opaque; + uint32_t ret = 0; + + addr >>= 2; + switch (addr) { + case R_SPI_RX: + s->regs[R_SPI_STATUS] &= ~S_RXFIFOFUL; + s->regs[R_SPI_STATUS] &= ~S_RXCHOVRF; + ret = fifo32_pop(&s->rx_fifo); + if (fifo32_is_empty(&s->rx_fifo)) { + s->regs[R_SPI_STATUS] |= S_RXFIFOEMP; + } + break; + + case R_SPI_MIS: + update_mis(s); + ret = s->regs[R_SPI_MIS]; + break; + + default: + if (addr < ARRAY_SIZE(s->regs)) { + ret = s->regs[addr]; + } else { + qemu_log_mask(LOG_GUEST_ERROR, + "%s: Bad offset 0x%" HWADDR_PRIx "\n", __func__, + addr * 4); + return ret; + } + break; + } + + DB_PRINT("addr=0x%" HWADDR_PRIx " = 0x%" PRIx32, addr * 4, ret); + spi_update_irq(s); + return ret; +} + +static void assert_cs(MSSSpiState *s) +{ + qemu_set_irq(s->cs_line, 0); +} + +static void deassert_cs(MSSSpiState *s) +{ + qemu_set_irq(s->cs_line, 1); +} + +static void spi_flush_txfifo(MSSSpiState *s) +{ + uint32_t tx; + uint32_t rx; + bool sps = !!(s->regs[R_SPI_CONTROL] & C_SPS); + + /* + * Chip Select(CS) is automatically controlled by this controller. + * If SPS bit is set in Control register then CS is asserted + * until all the frames set in frame count of Control register are + * transferred. If SPS is not set then CS pulses between frames. + * Note that Slave Select register specifies which of the CS line + * has to be controlled automatically by controller. Bits SS[7:1] are for + * masters in FPGA fabric since we model only Microcontroller subsystem + * of Smartfusion2 we control only one CS(SS[0]) line. + */ + while (!fifo32_is_empty(&s->tx_fifo) && s->frame_count) { + assert_cs(s); + + s->regs[R_SPI_STATUS] &= ~(S_TXDONE | S_RXRDY); + + tx = fifo32_pop(&s->tx_fifo); + DB_PRINT("data tx:0x%" PRIx32, tx); + rx = ssi_transfer(s->spi, tx); + DB_PRINT("data rx:0x%" PRIx32, rx); + + if (fifo32_num_used(&s->rx_fifo) == s->fifo_depth) { + s->regs[R_SPI_STATUS] |= S_RXCHOVRF; + s->regs[R_SPI_RIS] |= S_RXCHOVRF; + } else { + fifo32_push(&s->rx_fifo, rx); + s->regs[R_SPI_STATUS] &= ~S_RXFIFOEMP; + if (fifo32_num_used(&s->rx_fifo) == (s->fifo_depth - 1)) { + s->regs[R_SPI_STATUS] |= S_RXFIFOFULNXT; + } else if (fifo32_num_used(&s->rx_fifo) == s->fifo_depth) { + s->regs[R_SPI_STATUS] |= S_RXFIFOFUL; + } + } + s->frame_count--; + if (!sps) { + deassert_cs(s); + } + } + + if (!s->frame_count) { + s->frame_count = (s->regs[R_SPI_CONTROL] & FMCOUNT_MASK) >> + FMCOUNT_SHIFT; + deassert_cs(s); + s->regs[R_SPI_RIS] |= S_TXDONE | S_RXRDY; + s->regs[R_SPI_STATUS] |= S_TXDONE | S_RXRDY; + } +} + +static void spi_write(void *opaque, hwaddr addr, + uint64_t val64, unsigned int size) +{ + MSSSpiState *s = opaque; + uint32_t value = val64; + + DB_PRINT("addr=0x%" HWADDR_PRIx " =0x%" PRIx32, addr, value); + addr >>= 2; + + switch (addr) { + case R_SPI_TX: + /* adding to already full FIFO */ + if (fifo32_num_used(&s->tx_fifo) == s->fifo_depth) { + break; + } + s->regs[R_SPI_STATUS] &= ~S_TXFIFOEMP; + fifo32_push(&s->tx_fifo, value); + if (fifo32_num_used(&s->tx_fifo) == (s->fifo_depth - 1)) { + s->regs[R_SPI_STATUS] |= S_TXFIFOFULNXT; + } else if (fifo32_num_used(&s->tx_fifo) == s->fifo_depth) { + s->regs[R_SPI_STATUS] |= S_TXFIFOFUL; + } + if (s->enabled) { + spi_flush_txfifo(s); + } + break; + + case R_SPI_CONTROL: + s->regs[R_SPI_CONTROL] = value; + if (value & C_BIGFIFO) { + set_fifodepth(s); + } else { + s->fifo_depth = 4; + } + s->enabled = value & C_ENABLE; + s->frame_count = (value & FMCOUNT_MASK) >> FMCOUNT_SHIFT; + if (value & C_RESET) { + mss_spi_reset(DEVICE(s)); + } + break; + + case R_SPI_DFSIZE: + if (s->enabled) { + break; + } + s->regs[R_SPI_DFSIZE] = value; + break; + + case R_SPI_INTCLR: + s->regs[R_SPI_INTCLR] = value; + if (value & S_TXDONE) { + s->regs[R_SPI_RIS] &= ~S_TXDONE; + } + if (value & S_RXRDY) { + s->regs[R_SPI_RIS] &= ~S_RXRDY; + } + if (value & S_RXCHOVRF) { + s->regs[R_SPI_RIS] &= ~S_RXCHOVRF; + } + break; + + case R_SPI_MIS: + case R_SPI_STATUS: + case R_SPI_RIS: + qemu_log_mask(LOG_GUEST_ERROR, + "%s: Write to read only register 0x%" HWADDR_PRIx "\n", + __func__, addr * 4); + break; + + default: + if (addr < ARRAY_SIZE(s->regs)) { + s->regs[addr] = value; + } else { + qemu_log_mask(LOG_GUEST_ERROR, + "%s: Bad offset 0x%" HWADDR_PRIx "\n", __func__, + addr * 4); + } + break; + } + + spi_update_irq(s); +} + +static const MemoryRegionOps spi_ops = { + .read = spi_read, + .write = spi_write, + .endianness = DEVICE_NATIVE_ENDIAN, + .valid = { + .min_access_size = 1, + .max_access_size = 4 + } +}; + +static void mss_spi_realize(DeviceState *dev, Error **errp) +{ + MSSSpiState *s = MSS_SPI(dev); + SysBusDevice *sbd = SYS_BUS_DEVICE(dev); + + s->spi = ssi_create_bus(dev, "spi"); + + sysbus_init_irq(sbd, &s->irq); + ssi_auto_connect_slaves(dev, &s->cs_line, s->spi); + sysbus_init_irq(sbd, &s->cs_line); + + memory_region_init_io(&s->mmio, OBJECT(s), &spi_ops, s, + TYPE_MSS_SPI, R_SPI_MAX * 4); + sysbus_init_mmio(sbd, &s->mmio); + + fifo32_create(&s->tx_fifo, FIFO_CAPACITY); + fifo32_create(&s->rx_fifo, FIFO_CAPACITY); +} + +static const VMStateDescription vmstate_mss_spi = { + .name = TYPE_MSS_SPI, + .version_id = 1, + .minimum_version_id = 1, + .fields = (VMStateField[]) { + VMSTATE_FIFO32(tx_fifo, MSSSpiState), + VMSTATE_FIFO32(rx_fifo, MSSSpiState), + VMSTATE_UINT32_ARRAY(regs, MSSSpiState, R_SPI_MAX), + VMSTATE_END_OF_LIST() + } +}; + +static void mss_spi_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + + dc->realize = mss_spi_realize; + dc->reset = mss_spi_reset; + dc->vmsd = &vmstate_mss_spi; +} + +static const TypeInfo mss_spi_info = { + .name = TYPE_MSS_SPI, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(MSSSpiState), + .class_init = mss_spi_class_init, +}; + +static void mss_spi_register_types(void) +{ + type_register_static(&mss_spi_info); +} + +type_init(mss_spi_register_types) -- 2.7.4 From: Subbaraya Sundeep <sundeep.lkml@gmail.com> Smartfusion2 SoC has hardened Microcontroller subsystem and flash based FPGA fabric. This patch adds support for Microcontroller subsystem in the SoC. Signed-off-by: Subbaraya Sundeep <sundeep.lkml@gmail.com> Reviewed-by: Alistair Francis <alistair.francis@xilinx.com> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20170920201737.25723-5-f4bug@amsat.org [PMD: drop cpu_model to directly use cpu type, check m3clk non null] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/Makefile.objs | 1 + include/hw/arm/msf2-soc.h | 67 +++++++++++ hw/arm/msf2-soc.c | 238 ++++++++++++++++++++++++++++++++++++++++ default-configs/arm-softmmu.mak | 1 + 4 files changed, 307 insertions(+) create mode 100644 include/hw/arm/msf2-soc.h create mode 100644 hw/arm/msf2-soc.c diff --git a/hw/arm/Makefile.objs b/hw/arm/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Makefile.objs +++ b/hw/arm/Makefile.objs @@ -XXX,XX +XXX,XX @@ obj-$(CONFIG_FSL_IMX31) += fsl-imx31.o kzm.o obj-$(CONFIG_FSL_IMX6) += fsl-imx6.o sabrelite.o obj-$(CONFIG_ASPEED_SOC) += aspeed_soc.o aspeed.o obj-$(CONFIG_MPS2) += mps2.o +obj-$(CONFIG_MSF2) += msf2-soc.o diff --git a/include/hw/arm/msf2-soc.h b/include/hw/arm/msf2-soc.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/arm/msf2-soc.h @@ -XXX,XX +XXX,XX @@ +/* + * Microsemi Smartfusion2 SoC + * + * Copyright (c) 2017 Subbaraya Sundeep <sundeep.lkml@gmail.com> + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + */ + +#ifndef HW_ARM_MSF2_SOC_H +#define HW_ARM_MSF2_SOC_H + +#include "hw/arm/armv7m.h" +#include "hw/timer/mss-timer.h" +#include "hw/misc/msf2-sysreg.h" +#include "hw/ssi/mss-spi.h" + +#define TYPE_MSF2_SOC "msf2-soc" +#define MSF2_SOC(obj) OBJECT_CHECK(MSF2State, (obj), TYPE_MSF2_SOC) + +#define MSF2_NUM_SPIS 2 +#define MSF2_NUM_UARTS 2 + +/* + * System timer consists of two programmable 32-bit + * decrementing counters that generate individual interrupts to + * the Cortex-M3 processor + */ +#define MSF2_NUM_TIMERS 2 + +typedef struct MSF2State { + /*< private >*/ + SysBusDevice parent_obj; + /*< public >*/ + + ARMv7MState armv7m; + + char *cpu_type; + char *part_name; + uint64_t envm_size; + uint64_t esram_size; + + uint32_t m3clk; + uint8_t apb0div; + uint8_t apb1div; + + MSF2SysregState sysreg; + MSSTimerState timer; + MSSSpiState spi[MSF2_NUM_SPIS]; +} MSF2State; + +#endif diff --git a/hw/arm/msf2-soc.c b/hw/arm/msf2-soc.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/arm/msf2-soc.c @@ -XXX,XX +XXX,XX @@ +/* + * SmartFusion2 SoC emulation. + * + * Copyright (c) 2017 Subbaraya Sundeep <sundeep.lkml@gmail.com> + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + */ + +#include "qemu/osdep.h" +#include "qapi/error.h" +#include "qemu-common.h" +#include "hw/arm/arm.h" +#include "exec/address-spaces.h" +#include "hw/char/serial.h" +#include "hw/boards.h" +#include "sysemu/block-backend.h" +#include "qemu/cutils.h" +#include "hw/arm/msf2-soc.h" +#include "hw/misc/unimp.h" + +#define MSF2_TIMER_BASE 0x40004000 +#define MSF2_SYSREG_BASE 0x40038000 + +#define ENVM_BASE_ADDRESS 0x60000000 + +#define SRAM_BASE_ADDRESS 0x20000000 + +#define MSF2_ENVM_MAX_SIZE (512 * K_BYTE) + +/* + * eSRAM max size is 80k without SECDED(Single error correction and + * dual error detection) feature and 64k with SECDED. + * We do not support SECDED now. + */ +#define MSF2_ESRAM_MAX_SIZE (80 * K_BYTE) + +static const uint32_t spi_addr[MSF2_NUM_SPIS] = { 0x40001000 , 0x40011000 }; +static const uint32_t uart_addr[MSF2_NUM_UARTS] = { 0x40000000 , 0x40010000 }; + +static const int spi_irq[MSF2_NUM_SPIS] = { 2, 3 }; +static const int uart_irq[MSF2_NUM_UARTS] = { 10, 11 }; +static const int timer_irq[MSF2_NUM_TIMERS] = { 14, 15 }; + +static void m2sxxx_soc_initfn(Object *obj) +{ + MSF2State *s = MSF2_SOC(obj); + int i; + + object_initialize(&s->armv7m, sizeof(s->armv7m), TYPE_ARMV7M); + qdev_set_parent_bus(DEVICE(&s->armv7m), sysbus_get_default()); + + object_initialize(&s->sysreg, sizeof(s->sysreg), TYPE_MSF2_SYSREG); + qdev_set_parent_bus(DEVICE(&s->sysreg), sysbus_get_default()); + + object_initialize(&s->timer, sizeof(s->timer), TYPE_MSS_TIMER); + qdev_set_parent_bus(DEVICE(&s->timer), sysbus_get_default()); + + for (i = 0; i < MSF2_NUM_SPIS; i++) { + object_initialize(&s->spi[i], sizeof(s->spi[i]), + TYPE_MSS_SPI); + qdev_set_parent_bus(DEVICE(&s->spi[i]), sysbus_get_default()); + } +} + +static void m2sxxx_soc_realize(DeviceState *dev_soc, Error **errp) +{ + MSF2State *s = MSF2_SOC(dev_soc); + DeviceState *dev, *armv7m; + SysBusDevice *busdev; + Error *err = NULL; + int i; + + MemoryRegion *system_memory = get_system_memory(); + MemoryRegion *nvm = g_new(MemoryRegion, 1); + MemoryRegion *nvm_alias = g_new(MemoryRegion, 1); + MemoryRegion *sram = g_new(MemoryRegion, 1); + + memory_region_init_rom(nvm, NULL, "MSF2.eNVM", s->envm_size, + &error_fatal); + /* + * On power-on, the eNVM region 0x60000000 is automatically + * remapped to the Cortex-M3 processor executable region + * start address (0x0). We do not support remapping other eNVM, + * eSRAM and DDR regions by guest(via Sysreg) currently. + */ + memory_region_init_alias(nvm_alias, NULL, "MSF2.eNVM", + nvm, 0, s->envm_size); + + memory_region_add_subregion(system_memory, ENVM_BASE_ADDRESS, nvm); + memory_region_add_subregion(system_memory, 0, nvm_alias); + + memory_region_init_ram(sram, NULL, "MSF2.eSRAM", s->esram_size, + &error_fatal); + memory_region_add_subregion(system_memory, SRAM_BASE_ADDRESS, sram); + + armv7m = DEVICE(&s->armv7m); + qdev_prop_set_uint32(armv7m, "num-irq", 81); + qdev_prop_set_string(armv7m, "cpu-type", s->cpu_type); + object_property_set_link(OBJECT(&s->armv7m), OBJECT(get_system_memory()), + "memory", &error_abort); + object_property_set_bool(OBJECT(&s->armv7m), true, "realized", &err); + if (err != NULL) { + error_propagate(errp, err); + return; + } + + if (!s->m3clk) { + error_setg(errp, "Invalid m3clk value"); + error_append_hint(errp, "m3clk can not be zero\n"); + return; + } + system_clock_scale = NANOSECONDS_PER_SECOND / s->m3clk; + + for (i = 0; i < MSF2_NUM_UARTS; i++) { + if (serial_hds[i]) { + serial_mm_init(get_system_memory(), uart_addr[i], 2, + qdev_get_gpio_in(armv7m, uart_irq[i]), + 115200, serial_hds[i], DEVICE_NATIVE_ENDIAN); + } + } + + dev = DEVICE(&s->timer); + /* APB0 clock is the timer input clock */ + qdev_prop_set_uint32(dev, "clock-frequency", s->m3clk / s->apb0div); + object_property_set_bool(OBJECT(&s->timer), true, "realized", &err); + if (err != NULL) { + error_propagate(errp, err); + return; + } + busdev = SYS_BUS_DEVICE(dev); + sysbus_mmio_map(busdev, 0, MSF2_TIMER_BASE); + sysbus_connect_irq(busdev, 0, + qdev_get_gpio_in(armv7m, timer_irq[0])); + sysbus_connect_irq(busdev, 1, + qdev_get_gpio_in(armv7m, timer_irq[1])); + + dev = DEVICE(&s->sysreg); + qdev_prop_set_uint32(dev, "apb0divisor", s->apb0div); + qdev_prop_set_uint32(dev, "apb1divisor", s->apb1div); + object_property_set_bool(OBJECT(&s->sysreg), true, "realized", &err); + if (err != NULL) { + error_propagate(errp, err); + return; + } + busdev = SYS_BUS_DEVICE(dev); + sysbus_mmio_map(busdev, 0, MSF2_SYSREG_BASE); + + for (i = 0; i < MSF2_NUM_SPIS; i++) { + gchar *bus_name; + + object_property_set_bool(OBJECT(&s->spi[i]), true, "realized", &err); + if (err != NULL) { + error_propagate(errp, err); + return; + } + + sysbus_mmio_map(SYS_BUS_DEVICE(&s->spi[i]), 0, spi_addr[i]); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->spi[i]), 0, + qdev_get_gpio_in(armv7m, spi_irq[i])); + + /* Alias controller SPI bus to the SoC itself */ + bus_name = g_strdup_printf("spi%d", i); + object_property_add_alias(OBJECT(s), bus_name, + OBJECT(&s->spi[i]), "spi", + &error_abort); + g_free(bus_name); + } + + /* Below devices are not modelled yet. */ + create_unimplemented_device("i2c_0", 0x40002000, 0x1000); + create_unimplemented_device("dma", 0x40003000, 0x1000); + create_unimplemented_device("watchdog", 0x40005000, 0x1000); + create_unimplemented_device("i2c_1", 0x40012000, 0x1000); + create_unimplemented_device("gpio", 0x40013000, 0x1000); + create_unimplemented_device("hs-dma", 0x40014000, 0x1000); + create_unimplemented_device("can", 0x40015000, 0x1000); + create_unimplemented_device("rtc", 0x40017000, 0x1000); + create_unimplemented_device("apb_config", 0x40020000, 0x10000); + create_unimplemented_device("emac", 0x40041000, 0x1000); + create_unimplemented_device("usb", 0x40043000, 0x1000); +} + +static Property m2sxxx_soc_properties[] = { + /* + * part name specifies the type of SmartFusion2 device variant(this + * property is for information purpose only. + */ + DEFINE_PROP_STRING("cpu-type", MSF2State, cpu_type), + DEFINE_PROP_STRING("part-name", MSF2State, part_name), + DEFINE_PROP_UINT64("eNVM-size", MSF2State, envm_size, MSF2_ENVM_MAX_SIZE), + DEFINE_PROP_UINT64("eSRAM-size", MSF2State, esram_size, + MSF2_ESRAM_MAX_SIZE), + /* Libero GUI shows 100Mhz as default for clocks */ + DEFINE_PROP_UINT32("m3clk", MSF2State, m3clk, 100 * 1000000), + /* default divisors in Libero GUI */ + DEFINE_PROP_UINT8("apb0div", MSF2State, apb0div, 2), + DEFINE_PROP_UINT8("apb1div", MSF2State, apb1div, 2), + DEFINE_PROP_END_OF_LIST(), +}; + +static void m2sxxx_soc_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + + dc->realize = m2sxxx_soc_realize; + dc->props = m2sxxx_soc_properties; +} + +static const TypeInfo m2sxxx_soc_info = { + .name = TYPE_MSF2_SOC, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(MSF2State), + .instance_init = m2sxxx_soc_initfn, + .class_init = m2sxxx_soc_class_init, +}; + +static void m2sxxx_soc_types(void) +{ + type_register_static(&m2sxxx_soc_info); +} + +type_init(m2sxxx_soc_types) diff --git a/default-configs/arm-softmmu.mak b/default-configs/arm-softmmu.mak index XXXXXXX..XXXXXXX 100644 --- a/default-configs/arm-softmmu.mak +++ b/default-configs/arm-softmmu.mak @@ -XXX,XX +XXX,XX @@ CONFIG_ACPI=y CONFIG_SMBIOS=y CONFIG_ASPEED_SOC=y CONFIG_GPIO_KEY=y +CONFIG_MSF2=y -- 2.7.4 From: Subbaraya Sundeep <sundeep.lkml@gmail.com> Emulated Emcraft's Smartfusion2 System On Module starter kit. Signed-off-by: Subbaraya Sundeep <sundeep.lkml@gmail.com> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20170920201737.25723-6-f4bug@amsat.org [PMD: drop cpu_model to directly use cpu type] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/Makefile.objs | 2 +- hw/arm/msf2-som.c | 105 +++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 106 insertions(+), 1 deletion(-) create mode 100644 hw/arm/msf2-som.c diff --git a/hw/arm/Makefile.objs b/hw/arm/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Makefile.objs +++ b/hw/arm/Makefile.objs @@ -XXX,XX +XXX,XX @@ obj-$(CONFIG_FSL_IMX31) += fsl-imx31.o kzm.o obj-$(CONFIG_FSL_IMX6) += fsl-imx6.o sabrelite.o obj-$(CONFIG_ASPEED_SOC) += aspeed_soc.o aspeed.o obj-$(CONFIG_MPS2) += mps2.o -obj-$(CONFIG_MSF2) += msf2-soc.o +obj-$(CONFIG_MSF2) += msf2-soc.o msf2-som.o diff --git a/hw/arm/msf2-som.c b/hw/arm/msf2-som.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/arm/msf2-som.c @@ -XXX,XX +XXX,XX @@ +/* + * SmartFusion2 SOM starter kit(from Emcraft) emulation. + * + * Copyright (c) 2017 Subbaraya Sundeep <sundeep.lkml@gmail.com> + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + */ + +#include "qemu/osdep.h" +#include "qapi/error.h" +#include "qemu/error-report.h" +#include "hw/boards.h" +#include "hw/arm/arm.h" +#include "exec/address-spaces.h" +#include "qemu/cutils.h" +#include "hw/arm/msf2-soc.h" +#include "cpu.h" + +#define DDR_BASE_ADDRESS 0xA0000000 +#define DDR_SIZE (64 * M_BYTE) + +#define M2S010_ENVM_SIZE (256 * K_BYTE) +#define M2S010_ESRAM_SIZE (64 * K_BYTE) + +static void emcraft_sf2_s2s010_init(MachineState *machine) +{ + DeviceState *dev; + DeviceState *spi_flash; + MSF2State *soc; + MachineClass *mc = MACHINE_GET_CLASS(machine); + DriveInfo *dinfo = drive_get_next(IF_MTD); + qemu_irq cs_line; + SSIBus *spi_bus; + MemoryRegion *sysmem = get_system_memory(); + MemoryRegion *ddr = g_new(MemoryRegion, 1); + + if (strcmp(machine->cpu_type, mc->default_cpu_type) != 0) { + error_report("This board can only be used with CPU %s", + mc->default_cpu_type); + } + + memory_region_init_ram(ddr, NULL, "ddr-ram", DDR_SIZE, + &error_fatal); + memory_region_add_subregion(sysmem, DDR_BASE_ADDRESS, ddr); + + dev = qdev_create(NULL, TYPE_MSF2_SOC); + qdev_prop_set_string(dev, "part-name", "M2S010"); + qdev_prop_set_string(dev, "cpu-type", mc->default_cpu_type); + + qdev_prop_set_uint64(dev, "eNVM-size", M2S010_ENVM_SIZE); + qdev_prop_set_uint64(dev, "eSRAM-size", M2S010_ESRAM_SIZE); + + /* + * CPU clock and peripheral clocks(APB0, APB1)are configurable + * in Libero. CPU clock is divided by APB0 and APB1 divisors for + * peripherals. Emcraft's SoM kit comes with these settings by default. + */ + qdev_prop_set_uint32(dev, "m3clk", 142 * 1000000); + qdev_prop_set_uint32(dev, "apb0div", 2); + qdev_prop_set_uint32(dev, "apb1div", 2); + + object_property_set_bool(OBJECT(dev), true, "realized", &error_fatal); + + soc = MSF2_SOC(dev); + + /* Attach SPI flash to SPI0 controller */ + spi_bus = (SSIBus *)qdev_get_child_bus(dev, "spi0"); + spi_flash = ssi_create_slave_no_init(spi_bus, "s25sl12801"); + qdev_prop_set_uint8(spi_flash, "spansion-cr2nv", 1); + if (dinfo) { + qdev_prop_set_drive(spi_flash, "drive", blk_by_legacy_dinfo(dinfo), + &error_fatal); + } + qdev_init_nofail(spi_flash); + cs_line = qdev_get_gpio_in_named(spi_flash, SSI_GPIO_CS, 0); + sysbus_connect_irq(SYS_BUS_DEVICE(&soc->spi[0]), 1, cs_line); + + armv7m_load_kernel(ARM_CPU(first_cpu), machine->kernel_filename, + soc->envm_size); +} + +static void emcraft_sf2_machine_init(MachineClass *mc) +{ + mc->desc = "SmartFusion2 SOM kit from Emcraft (M2S010)"; + mc->init = emcraft_sf2_s2s010_init; + mc->default_cpu_type = ARM_CPU_TYPE_NAME("cortex-m3"); +} + +DEFINE_MACHINE("emcraft-sf2", emcraft_sf2_machine_init) -- 2.7.4 I don't have anything else queued up at the moment, so this is just Richard's SME patches. -- PMM The following changes since commit 63b38f6c85acd312c2cab68554abf33adf4ee2b3: Merge tag 'pull-target-arm-20220707' of https://git.linaro.org/people/pmaydell/qemu-arm into staging (2022-07-08 06:17:11 +0530) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20220711 for you to fetch changes up to f9982ceaf26df27d15547a3a7990a95019e9e3a8: linux-user/aarch64: Add SME related hwcap entries (2022-07-11 13:43:52 +0100) ---------------------------------------------------------------- target-arm: * Implement SME emulation, for both system and linux-user ---------------------------------------------------------------- Richard Henderson (45): target/arm: Handle SME in aarch64_cpu_dump_state target/arm: Add infrastructure for disas_sme target/arm: Trap non-streaming usage when Streaming SVE is active target/arm: Mark ADR as non-streaming target/arm: Mark RDFFR, WRFFR, SETFFR as non-streaming target/arm: Mark BDEP, BEXT, BGRP, COMPACT, FEXPA, FTSSEL as non-streaming target/arm: Mark PMULL, FMMLA as non-streaming target/arm: Mark FTSMUL, FTMAD, FADDA as non-streaming target/arm: Mark SMMLA, UMMLA, USMMLA as non-streaming target/arm: Mark string/histo/crypto as non-streaming target/arm: Mark gather/scatter load/store as non-streaming target/arm: Mark gather prefetch as non-streaming target/arm: Mark LDFF1 and LDNF1 as non-streaming target/arm: Mark LD1RO as non-streaming target/arm: Add SME enablement checks target/arm: Handle SME in sve_access_check target/arm: Implement SME RDSVL, ADDSVL, ADDSPL target/arm: Implement SME ZERO target/arm: Implement SME MOVA target/arm: Implement SME LD1, ST1 target/arm: Export unpredicated ld/st from translate-sve.c target/arm: Implement SME LDR, STR target/arm: Implement SME ADDHA, ADDVA target/arm: Implement FMOPA, FMOPS (non-widening) target/arm: Implement BFMOPA, BFMOPS target/arm: Implement FMOPA, FMOPS (widening) target/arm: Implement SME integer outer product target/arm: Implement PSEL target/arm: Implement REVD target/arm: Implement SCLAMP, UCLAMP target/arm: Reset streaming sve state on exception boundaries target/arm: Enable SME for -cpu max linux-user/aarch64: Clear tpidr2_el0 if CLONE_SETTLS linux-user/aarch64: Reset PSTATE.SM on syscalls linux-user/aarch64: Add SM bit to SVE signal context linux-user/aarch64: Tidy target_restore_sigframe error return linux-user/aarch64: Do not allow duplicate or short sve records linux-user/aarch64: Verify extra record lock succeeded linux-user/aarch64: Move sve record checks into restore linux-user/aarch64: Implement SME signal handling linux-user: Rename sve prctls linux-user/aarch64: Implement PR_SME_GET_VL, PR_SME_SET_VL target/arm: Only set ZEN in reset if SVE present target/arm: Enable SME for user-only linux-user/aarch64: Add SME related hwcap entries docs/system/arm/emulation.rst | 4 + linux-user/aarch64/target_cpu.h | 5 +- linux-user/aarch64/target_prctl.h | 62 +- target/arm/cpu.h | 7 + target/arm/helper-sme.h | 126 ++++ target/arm/helper-sve.h | 4 + target/arm/helper.h | 18 + target/arm/translate-a64.h | 45 ++ target/arm/translate.h | 16 + target/arm/sme-fa64.decode | 60 ++ target/arm/sme.decode | 88 +++ target/arm/sve.decode | 41 +- linux-user/aarch64/cpu_loop.c | 9 + linux-user/aarch64/signal.c | 243 ++++++-- linux-user/elfload.c | 20 + linux-user/syscall.c | 28 +- target/arm/cpu.c | 35 +- target/arm/cpu64.c | 11 + target/arm/helper.c | 56 +- target/arm/sme_helper.c | 1140 +++++++++++++++++++++++++++++++++++++ target/arm/sve_helper.c | 28 + target/arm/translate-a64.c | 103 +++- target/arm/translate-sme.c | 373 ++++++++++++ target/arm/translate-sve.c | 393 ++++++++++--- target/arm/translate-vfp.c | 12 + target/arm/translate.c | 2 + target/arm/vec_helper.c | 24 + target/arm/meson.build | 3 + 28 files changed, 2821 insertions(+), 135 deletions(-) create mode 100644 target/arm/sme-fa64.decode create mode 100644 target/arm/sme.decode create mode 100644 target/arm/translate-sme.c From: Richard Henderson <richard.henderson@linaro.org> Dump SVCR, plus use the correct access check for Streaming Mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-2-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ static void aarch64_cpu_dump_state(CPUState *cs, FILE *f, int flags) int i; int el = arm_current_el(env); const char *ns_status; + bool sve; qemu_fprintf(f, " PC=%016" PRIx64 " ", env->pc); for (i = 0; i < 32; i++) { @@ -XXX,XX +XXX,XX @@ static void aarch64_cpu_dump_state(CPUState *cs, FILE *f, int flags) el, psr & PSTATE_SP ? 'h' : 't'); + if (cpu_isar_feature(aa64_sme, cpu)) { + qemu_fprintf(f, " SVCR=%08" PRIx64 " %c%c", + env->svcr, + (FIELD_EX64(env->svcr, SVCR, ZA) ? 'Z' : '-'), + (FIELD_EX64(env->svcr, SVCR, SM) ? 'S' : '-')); + } if (cpu_isar_feature(aa64_bti, cpu)) { qemu_fprintf(f, " BTYPE=%d", (psr & PSTATE_BTYPE) >> 10); } @@ -XXX,XX +XXX,XX @@ static void aarch64_cpu_dump_state(CPUState *cs, FILE *f, int flags) qemu_fprintf(f, " FPCR=%08x FPSR=%08x\n", vfp_get_fpcr(env), vfp_get_fpsr(env)); - if (cpu_isar_feature(aa64_sve, cpu) && sve_exception_el(env, el) == 0) { + if (cpu_isar_feature(aa64_sme, cpu) && FIELD_EX64(env->svcr, SVCR, SM)) { + sve = sme_exception_el(env, el) == 0; + } else if (cpu_isar_feature(aa64_sve, cpu)) { + sve = sve_exception_el(env, el) == 0; + } else { + sve = false; + } + + if (sve) { int j, zcr_len = sve_vqm1_for_el(env, el); for (i = 0; i <= FFR_PRED_NUM; i++) { -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This includes the build rules for the decoder, and the new file for translation, but excludes any instructions. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-3-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.h | 1 + target/arm/sme.decode | 20 ++++++++++++++++++++ target/arm/translate-a64.c | 7 ++++++- target/arm/translate-sme.c | 35 +++++++++++++++++++++++++++++++++++ target/arm/meson.build | 2 ++ 5 files changed, 64 insertions(+), 1 deletion(-) create mode 100644 target/arm/sme.decode create mode 100644 target/arm/translate-sme.c diff --git a/target/arm/translate-a64.h b/target/arm/translate-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.h +++ b/target/arm/translate-a64.h @@ -XXX,XX +XXX,XX @@ static inline int pred_gvec_reg_size(DisasContext *s) } bool disas_sve(DisasContext *, uint32_t); +bool disas_sme(DisasContext *, uint32_t); void gen_gvec_rax1(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); diff --git a/target/arm/sme.decode b/target/arm/sme.decode new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ +# AArch64 SME instruction descriptions +# +# Copyright (c) 2022 Linaro, Ltd +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library; if not, see <http://www.gnu.org/licenses/>. + +# +# This file is processed by scripts/decodetree.py +# 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 aarch64_tr_translate_insn(DisasContextBase *dcbase, CPUState *cpu) } switch (extract32(insn, 25, 4)) { - case 0x0: case 0x1: case 0x3: /* UNALLOCATED */ + case 0x0: + if (!extract32(insn, 31, 1) || !disas_sme(s, insn)) { + unallocated_encoding(s); + } + break; + case 0x1: case 0x3: /* UNALLOCATED */ unallocated_encoding(s); break; case 0x2: diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ +/* + * AArch64 SME translation + * + * Copyright (c) 2022 Linaro, Ltd + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, see <http://www.gnu.org/licenses/>. + */ + +#include "qemu/osdep.h" +#include "cpu.h" +#include "tcg/tcg-op.h" +#include "tcg/tcg-op-gvec.h" +#include "tcg/tcg-gvec-desc.h" +#include "translate.h" +#include "exec/helper-gen.h" +#include "translate-a64.h" +#include "fpu/softfloat.h" + + +/* + * Include the generated decoder. + */ + +#include "decode-sme.c.inc" diff --git a/target/arm/meson.build b/target/arm/meson.build index XXXXXXX..XXXXXXX 100644 --- a/target/arm/meson.build +++ b/target/arm/meson.build @@ -XXX,XX +XXX,XX @@ gen = [ decodetree.process('sve.decode', extra_args: '--decode=disas_sve'), + decodetree.process('sme.decode', extra_args: '--decode=disas_sme'), decodetree.process('neon-shared.decode', extra_args: '--decode=disas_neon_shared'), decodetree.process('neon-dp.decode', extra_args: '--decode=disas_neon_dp'), decodetree.process('neon-ls.decode', extra_args: '--decode=disas_neon_ls'), @@ -XXX,XX +XXX,XX @@ arm_ss.add(when: 'TARGET_AARCH64', if_true: files( 'sme_helper.c', 'translate-a64.c', 'translate-sve.c', + 'translate-sme.c', )) arm_softmmu_ss = ss.source_set() -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This new behaviour is in the ARM pseudocode function AArch64.CheckFPAdvSIMDEnabled, which applies to AArch32 via AArch32.CheckAdvSIMDOrFPEnabled when the EL to which the trap would be delivered is in AArch64 mode. Given that ARMv9 drops support for AArch32 outside EL0, the trap EL detection ought to be trivially true, but the pseudocode still contains a number of conditions, and QEMU has not yet committed to dropping A32 support for EL[12] when v9 features are present. Since the computation of SME_TRAP_NONSTREAMING is necessarily different for the two modes, we might as well preserve bits within TBFLAG_ANY and allocate separate bits within TBFLAG_A32 and TBFLAG_A64 instead. Note that DDI0616A.a has typos for bits [22:21] of LD1RO in the table of instructions illegal in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-4-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 7 +++ target/arm/translate.h | 4 ++ target/arm/sme-fa64.decode | 90 ++++++++++++++++++++++++++++++++++++++ target/arm/helper.c | 41 +++++++++++++++++ target/arm/translate-a64.c | 40 ++++++++++++++++- target/arm/translate-vfp.c | 12 +++++ target/arm/translate.c | 2 + target/arm/meson.build | 1 + 8 files changed, 195 insertions(+), 2 deletions(-) create mode 100644 target/arm/sme-fa64.decode 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(TBFLAG_A32, HSTR_ACTIVE, 9, 1) * the same thing as the current security state of the processor! */ FIELD(TBFLAG_A32, NS, 10, 1) +/* + * Indicates that SME Streaming mode is active, and SMCR_ELx.FA64 is not. + * This requires an SME trap from AArch32 mode when using NEON. + */ +FIELD(TBFLAG_A32, SME_TRAP_NONSTREAMING, 11, 1) /* * Bit usage when in AArch32 state, for M-profile only. @@ -XXX,XX +XXX,XX @@ FIELD(TBFLAG_A64, SMEEXC_EL, 20, 2) FIELD(TBFLAG_A64, PSTATE_SM, 22, 1) FIELD(TBFLAG_A64, PSTATE_ZA, 23, 1) FIELD(TBFLAG_A64, SVL, 24, 4) +/* Indicates that SME Streaming mode is active, and SMCR_ELx.FA64 is not. */ +FIELD(TBFLAG_A64, SME_TRAP_NONSTREAMING, 28, 1) /* * Helpers for using the above. 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 { bool pstate_sm; /* True if PSTATE.ZA is set. */ bool pstate_za; + /* True if non-streaming insns should raise an SME Streaming exception. */ + bool sme_trap_nonstreaming; + /* True if the current instruction is non-streaming. */ + bool is_nonstreaming; /* True if MVE insns are definitely not predicated by VPR or LTPSIZE */ bool mve_no_pred; /* diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ +# AArch64 SME allowed instruction decoding +# +# Copyright (c) 2022 Linaro, Ltd +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library; if not, see <http://www.gnu.org/licenses/>. + +# +# This file is processed by scripts/decodetree.py +# + +# These patterns are taken from Appendix E1.1 of DDI0616 A.a, +# Arm Architecture Reference Manual Supplement, +# The Scalable Matrix Extension (SME), for Armv9-A + +{ + [ + OK 0-00 1110 0000 0001 0010 11-- ---- ---- # SMOV W|Xd,Vn.B[0] + OK 0-00 1110 0000 0010 0010 11-- ---- ---- # SMOV W|Xd,Vn.H[0] + OK 0100 1110 0000 0100 0010 11-- ---- ---- # SMOV Xd,Vn.S[0] + OK 0000 1110 0000 0001 0011 11-- ---- ---- # UMOV Wd,Vn.B[0] + OK 0000 1110 0000 0010 0011 11-- ---- ---- # UMOV Wd,Vn.H[0] + OK 0000 1110 0000 0100 0011 11-- ---- ---- # UMOV Wd,Vn.S[0] + OK 0100 1110 0000 1000 0011 11-- ---- ---- # UMOV Xd,Vn.D[0] + ] + FAIL 0--0 111- ---- ---- ---- ---- ---- ---- # Advanced SIMD vector operations +} + +{ + [ + OK 0101 1110 --1- ---- 11-1 11-- ---- ---- # FMULX/FRECPS/FRSQRTS (scalar) + OK 0101 1110 -10- ---- 00-1 11-- ---- ---- # FMULX/FRECPS/FRSQRTS (scalar, FP16) + OK 01-1 1110 1-10 0001 11-1 10-- ---- ---- # FRECPE/FRSQRTE/FRECPX (scalar) + OK 01-1 1110 1111 1001 11-1 10-- ---- ---- # FRECPE/FRSQRTE/FRECPX (scalar, FP16) + ] + FAIL 01-1 111- ---- ---- ---- ---- ---- ---- # Advanced SIMD single-element operations +} + +FAIL 0-00 110- ---- ---- ---- ---- ---- ---- # Advanced SIMD structure load/store +FAIL 1100 1110 ---- ---- ---- ---- ---- ---- # Advanced SIMD cryptography extensions +FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS + +# These are the "avoidance of doubt" final table of Illegal Advanced SIMD instructions +# We don't actually need to include these, as the default is OK. +# -001 111- ---- ---- ---- ---- ---- ---- # Scalar floating-point operations +# --10 110- ---- ---- ---- ---- ---- ---- # Load/store pair of FP registers +# --01 1100 ---- ---- ---- ---- ---- ---- # Load FP register (PC-relative literal) +# --11 1100 --0- ---- ---- ---- ---- ---- # Load/store FP register (unscaled imm) +# --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) +# --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) + +FAIL 0000 0100 --1- ---- 1010 ---- ---- ---- # ADR +FAIL 0000 0100 --1- ---- 1011 -0-- ---- ---- # FTSSEL, FEXPA +FAIL 0000 0101 --10 0001 100- ---- ---- ---- # COMPACT +FAIL 0010 0101 --01 100- 1111 000- ---0 ---- # RDFFR, RDFFRS +FAIL 0010 0101 --10 1--- 1001 ---- ---- ---- # WRFFR, SETFFR +FAIL 0100 0101 --0- ---- 1011 ---- ---- ---- # BDEP, BEXT, BGRP +FAIL 0100 0101 000- ---- 0110 1--- ---- ---- # PMULLB, PMULLT (128b result) +FAIL 0110 0100 --1- ---- 1110 01-- ---- ---- # FMMLA, BFMMLA +FAIL 0110 0101 --0- ---- 0000 11-- ---- ---- # FTSMUL +FAIL 0110 0101 --01 0--- 100- ---- ---- ---- # FTMAD +FAIL 0110 0101 --01 1--- 001- ---- ---- ---- # FADDA +FAIL 0100 0101 --0- ---- 1001 10-- ---- ---- # SMMLA, UMMLA, USMMLA +FAIL 0100 0101 --1- ---- 1--- ---- ---- ---- # SVE2 string/histo/crypto instructions +FAIL 1000 010- -00- ---- 10-- ---- ---- ---- # SVE2 32-bit gather NT load (vector+scalar) +FAIL 1000 010- -00- ---- 111- ---- ---- ---- # SVE 32-bit gather prefetch (vector+imm) +FAIL 1000 0100 0-1- ---- 0--- ---- ---- ---- # SVE 32-bit gather prefetch (scalar+vector) +FAIL 1000 010- -01- ---- 1--- ---- ---- ---- # SVE 32-bit gather load (vector+imm) +FAIL 1000 0100 0-0- ---- 0--- ---- ---- ---- # SVE 32-bit gather load byte (scalar+vector) +FAIL 1000 0100 1--- ---- 0--- ---- ---- ---- # SVE 32-bit gather load half (scalar+vector) +FAIL 1000 0101 0--- ---- 0--- ---- ---- ---- # SVE 32-bit gather load word (scalar+vector) +FAIL 1010 010- ---- ---- 011- ---- ---- ---- # SVE contiguous FF load (scalar+scalar) +FAIL 1010 010- ---1 ---- 101- ---- ---- ---- # SVE contiguous NF load (scalar+imm) +FAIL 1010 010- -01- ---- 000- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+scalar) +FAIL 1010 010- -010 ---- 001- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+imm) +FAIL 1100 010- ---- ---- ---- ---- ---- ---- # SVE 64-bit gather load/prefetch +FAIL 1110 010- -00- ---- 001- ---- ---- ---- # SVE2 64-bit scatter NT store (vector+scalar) +FAIL 1110 010- -10- ---- 001- ---- ---- ---- # SVE2 32-bit scatter NT store (vector+scalar) +FAIL 1110 010- ---- ---- 1-0- ---- ---- ---- # SVE scatter store (scalar+32-bit vector) +FAIL 1110 010- ---- ---- 101- ---- ---- ---- # SVE scatter store (misc) 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 @@ int sme_exception_el(CPUARMState *env, int el) return 0; } +/* This corresponds to the ARM pseudocode function IsFullA64Enabled(). */ +static bool sme_fa64(CPUARMState *env, int el) +{ + if (!cpu_isar_feature(aa64_sme_fa64, env_archcpu(env))) { + return false; + } + + if (el <= 1 && !el_is_in_host(env, el)) { + if (!FIELD_EX64(env->vfp.smcr_el[1], SMCR, FA64)) { + return false; + } + } + if (el <= 2 && arm_is_el2_enabled(env)) { + if (!FIELD_EX64(env->vfp.smcr_el[2], SMCR, FA64)) { + return false; + } + } + if (arm_feature(env, ARM_FEATURE_EL3)) { + if (!FIELD_EX64(env->vfp.smcr_el[3], SMCR, FA64)) { + return false; + } + } + + return true; +} + /* * Given that SVE is enabled, return the vector length for EL. */ @@ -XXX,XX +XXX,XX @@ static CPUARMTBFlags rebuild_hflags_a32(CPUARMState *env, int fp_el, DP_TBFLAG_ANY(flags, PSTATE__IL, 1); } + /* + * The SME exception we are testing for is raised via + * AArch64.CheckFPAdvSIMDEnabled(), as called from + * AArch32.CheckAdvSIMDOrFPEnabled(). + */ + if (el == 0 + && FIELD_EX64(env->svcr, SVCR, SM) + && (!arm_is_el2_enabled(env) + || (arm_el_is_aa64(env, 2) && !(env->cp15.hcr_el2 & HCR_TGE))) + && arm_el_is_aa64(env, 1) + && !sme_fa64(env, el)) { + DP_TBFLAG_A32(flags, SME_TRAP_NONSTREAMING, 1); + } + return rebuild_hflags_common_32(env, fp_el, mmu_idx, flags); } @@ -XXX,XX +XXX,XX @@ static CPUARMTBFlags rebuild_hflags_a64(CPUARMState *env, int el, int fp_el, } if (FIELD_EX64(env->svcr, SVCR, SM)) { DP_TBFLAG_A64(flags, PSTATE_SM, 1); + DP_TBFLAG_A64(flags, SME_TRAP_NONSTREAMING, !sme_fa64(env, el)); } DP_TBFLAG_A64(flags, PSTATE_ZA, FIELD_EX64(env->svcr, SVCR, ZA)); } 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 bool fp_access_check(DisasContext *s) +static bool fp_access_check_only(DisasContext *s) { if (s->fp_excp_el) { assert(!s->fp_access_checked); @@ -XXX,XX +XXX,XX @@ static bool fp_access_check(DisasContext *s) return true; } +static bool fp_access_check(DisasContext *s) +{ + if (!fp_access_check_only(s)) { + return false; + } + if (s->sme_trap_nonstreaming && s->is_nonstreaming) { + gen_exception_insn(s, s->pc_curr, EXCP_UDEF, + syn_smetrap(SME_ET_Streaming, false)); + return false; + } + return true; +} + /* Check that SVE access is enabled. If it is, return true. * If not, emit code to generate an appropriate exception and return false. */ @@ -XXX,XX +XXX,XX @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread, default: g_assert_not_reached(); } - if ((ri->type & ARM_CP_FPU) && !fp_access_check(s)) { + if ((ri->type & ARM_CP_FPU) && !fp_access_check_only(s)) { return; } else if ((ri->type & ARM_CP_SVE) && !sve_access_check(s)) { return; @@ -XXX,XX +XXX,XX @@ static void disas_data_proc_simd_fp(DisasContext *s, uint32_t insn) } } +/* + * Include the generated SME FA64 decoder. + */ + +#include "decode-sme-fa64.c.inc" + +static bool trans_OK(DisasContext *s, arg_OK *a) +{ + return true; +} + +static bool trans_FAIL(DisasContext *s, arg_OK *a) +{ + s->is_nonstreaming = true; + return true; +} + /** * is_guarded_page: * @env: The cpu environment @@ -XXX,XX +XXX,XX @@ static void aarch64_tr_init_disas_context(DisasContextBase *dcbase, dc->mte_active[1] = EX_TBFLAG_A64(tb_flags, MTE0_ACTIVE); dc->pstate_sm = EX_TBFLAG_A64(tb_flags, PSTATE_SM); dc->pstate_za = EX_TBFLAG_A64(tb_flags, PSTATE_ZA); + dc->sme_trap_nonstreaming = EX_TBFLAG_A64(tb_flags, SME_TRAP_NONSTREAMING); dc->vec_len = 0; dc->vec_stride = 0; dc->cp_regs = arm_cpu->cp_regs; @@ -XXX,XX +XXX,XX @@ static void aarch64_tr_translate_insn(DisasContextBase *dcbase, CPUState *cpu) } } + s->is_nonstreaming = false; + if (s->sme_trap_nonstreaming) { + disas_sme_fa64(s, insn); + } + switch (extract32(insn, 25, 4)) { case 0x0: if (!extract32(insn, 31, 1) || !disas_sme(s, insn)) { diff --git a/target/arm/translate-vfp.c b/target/arm/translate-vfp.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-vfp.c +++ b/target/arm/translate-vfp.c @@ -XXX,XX +XXX,XX @@ static bool vfp_access_check_a(DisasContext *s, bool ignore_vfp_enabled) return false; } + /* + * Note that rebuild_hflags_a32 has already accounted for being in EL0 + * and the higher EL in A64 mode, etc. Unlike A64 mode, there do not + * appear to be any insns which touch VFP which are allowed. + */ + if (s->sme_trap_nonstreaming) { + gen_exception_insn(s, s->pc_curr, EXCP_UDEF, + syn_smetrap(SME_ET_Streaming, + s->base.pc_next - s->pc_curr == 2)); + return false; + } + if (!s->vfp_enabled && !ignore_vfp_enabled) { assert(!arm_dc_feature(s, ARM_FEATURE_M)); unallocated_encoding(s); 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->vec_len = EX_TBFLAG_A32(tb_flags, VECLEN); dc->vec_stride = EX_TBFLAG_A32(tb_flags, VECSTRIDE); } + dc->sme_trap_nonstreaming = + EX_TBFLAG_A32(tb_flags, SME_TRAP_NONSTREAMING); } dc->cp_regs = cpu->cp_regs; dc->features = env->features; diff --git a/target/arm/meson.build b/target/arm/meson.build index XXXXXXX..XXXXXXX 100644 --- a/target/arm/meson.build +++ b/target/arm/meson.build @@ -XXX,XX +XXX,XX @@ gen = [ decodetree.process('sve.decode', extra_args: '--decode=disas_sve'), decodetree.process('sme.decode', extra_args: '--decode=disas_sme'), + decodetree.process('sme-fa64.decode', extra_args: '--static-decode=disas_sme_fa64'), decodetree.process('neon-shared.decode', extra_args: '--decode=disas_neon_shared'), decodetree.process('neon-dp.decode', extra_args: '--decode=disas_neon_dp'), decodetree.process('neon-ls.decode', extra_args: '--decode=disas_neon_ls'), -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark ADR as a non-streaming instruction, which should trap if full a64 support is not enabled in streaming mode. Removing entries from sme-fa64.decode is an easy way to see what remains to be done. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-5-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.h | 7 +++++++ target/arm/sme-fa64.decode | 1 - target/arm/translate-sve.c | 8 ++++---- 3 files changed, 11 insertions(+), 5 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 @@ uint64_t asimd_imm_const(uint32_t imm, int cmode, int op); static bool trans_##NAME(DisasContext *s, arg_##NAME *a) \ { return dc_isar_feature(FEAT, s) && FUNC(s, __VA_ARGS__); } +#define TRANS_FEAT_NONSTREAMING(NAME, FEAT, FUNC, ...) \ + static bool trans_##NAME(DisasContext *s, arg_##NAME *a) \ + { \ + s->is_nonstreaming = true; \ + return dc_isar_feature(FEAT, s) && FUNC(s, __VA_ARGS__); \ + } + #endif /* TARGET_ARM_TRANSLATE_H */ diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 0000 0100 --1- ---- 1010 ---- ---- ---- # ADR FAIL 0000 0100 --1- ---- 1011 -0-- ---- ---- # FTSSEL, FEXPA FAIL 0000 0101 --10 0001 100- ---- ---- ---- # COMPACT FAIL 0010 0101 --01 100- 1111 000- ---0 ---- # RDFFR, RDFFRS 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 do_adr(DisasContext *s, arg_rrri *a, gen_helper_gvec_3 *fn) return gen_gvec_ool_zzz(s, fn, a->rd, a->rn, a->rm, a->imm); } -TRANS_FEAT(ADR_p32, aa64_sve, do_adr, a, gen_helper_sve_adr_p32) -TRANS_FEAT(ADR_p64, aa64_sve, do_adr, a, gen_helper_sve_adr_p64) -TRANS_FEAT(ADR_s32, aa64_sve, do_adr, a, gen_helper_sve_adr_s32) -TRANS_FEAT(ADR_u32, aa64_sve, do_adr, a, gen_helper_sve_adr_u32) +TRANS_FEAT_NONSTREAMING(ADR_p32, aa64_sve, do_adr, a, gen_helper_sve_adr_p32) +TRANS_FEAT_NONSTREAMING(ADR_p64, aa64_sve, do_adr, a, gen_helper_sve_adr_p64) +TRANS_FEAT_NONSTREAMING(ADR_s32, aa64_sve, do_adr, a, gen_helper_sve_adr_s32) +TRANS_FEAT_NONSTREAMING(ADR_u32, aa64_sve, do_adr, a, gen_helper_sve_adr_u32) /* *** SVE Integer Misc - Unpredicated Group -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-6-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 2 -- target/arm/translate-sve.c | 9 ++++++--- 2 files changed, 6 insertions(+), 5 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS FAIL 0000 0100 --1- ---- 1011 -0-- ---- ---- # FTSSEL, FEXPA FAIL 0000 0101 --10 0001 100- ---- ---- ---- # COMPACT -FAIL 0010 0101 --01 100- 1111 000- ---0 ---- # RDFFR, RDFFRS -FAIL 0010 0101 --10 1--- 1001 ---- ---- ---- # WRFFR, SETFFR FAIL 0100 0101 --0- ---- 1011 ---- ---- ---- # BDEP, BEXT, BGRP FAIL 0100 0101 000- ---- 0110 1--- ---- ---- # PMULLB, PMULLT (128b result) FAIL 0110 0100 --1- ---- 1110 01-- ---- ---- # FMMLA, BFMMLA 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 do_predset(DisasContext *s, int esz, int rd, int pat, bool setflag) TRANS_FEAT(PTRUE, aa64_sve, do_predset, a->esz, a->rd, a->pat, a->s) /* Note pat == 31 is #all, to set all elements. */ -TRANS_FEAT(SETFFR, aa64_sve, do_predset, 0, FFR_PRED_NUM, 31, false) +TRANS_FEAT_NONSTREAMING(SETFFR, aa64_sve, + do_predset, 0, FFR_PRED_NUM, 31, false) /* Note pat == 32 is #unimp, to set no elements. */ TRANS_FEAT(PFALSE, aa64_sve, do_predset, 0, a->rd, 32, false) @@ -XXX,XX +XXX,XX @@ static bool trans_RDFFR_p(DisasContext *s, arg_RDFFR_p *a) .rd = a->rd, .pg = a->pg, .s = a->s, .rn = FFR_PRED_NUM, .rm = FFR_PRED_NUM, }; + + s->is_nonstreaming = true; return trans_AND_pppp(s, &alt_a); } -TRANS_FEAT(RDFFR, aa64_sve, do_mov_p, a->rd, FFR_PRED_NUM) -TRANS_FEAT(WRFFR, aa64_sve, do_mov_p, FFR_PRED_NUM, a->rn) +TRANS_FEAT_NONSTREAMING(RDFFR, aa64_sve, do_mov_p, a->rd, FFR_PRED_NUM) +TRANS_FEAT_NONSTREAMING(WRFFR, aa64_sve, do_mov_p, FFR_PRED_NUM, a->rn) static bool do_pfirst_pnext(DisasContext *s, arg_rr_esz *a, void (*gen_fn)(TCGv_i32, TCGv_ptr, -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-7-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 3 --- target/arm/translate-sve.c | 22 ++++++++++++---------- 2 files changed, 12 insertions(+), 13 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 0000 0100 --1- ---- 1011 -0-- ---- ---- # FTSSEL, FEXPA -FAIL 0000 0101 --10 0001 100- ---- ---- ---- # COMPACT -FAIL 0100 0101 --0- ---- 1011 ---- ---- ---- # BDEP, BEXT, BGRP FAIL 0100 0101 000- ---- 0110 1--- ---- ---- # PMULLB, PMULLT (128b result) FAIL 0110 0100 --1- ---- 1110 01-- ---- ---- # FMMLA, BFMMLA FAIL 0110 0101 --0- ---- 0000 11-- ---- ---- # FTSMUL 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_2 * const fexpa_fns[4] = { NULL, gen_helper_sve_fexpa_h, gen_helper_sve_fexpa_s, gen_helper_sve_fexpa_d, }; -TRANS_FEAT(FEXPA, aa64_sve, gen_gvec_ool_zz, - fexpa_fns[a->esz], a->rd, a->rn, 0) +TRANS_FEAT_NONSTREAMING(FEXPA, aa64_sve, gen_gvec_ool_zz, + fexpa_fns[a->esz], a->rd, a->rn, 0) static gen_helper_gvec_3 * const ftssel_fns[4] = { NULL, gen_helper_sve_ftssel_h, gen_helper_sve_ftssel_s, gen_helper_sve_ftssel_d, }; -TRANS_FEAT(FTSSEL, aa64_sve, gen_gvec_ool_arg_zzz, ftssel_fns[a->esz], a, 0) +TRANS_FEAT_NONSTREAMING(FTSSEL, aa64_sve, gen_gvec_ool_arg_zzz, + ftssel_fns[a->esz], a, 0) /* *** SVE Predicate Logical Operations Group @@ -XXX,XX +XXX,XX @@ TRANS_FEAT(TRN2_q, aa64_sve_f64mm, gen_gvec_ool_arg_zzz, static gen_helper_gvec_3 * const compact_fns[4] = { NULL, NULL, gen_helper_sve_compact_s, gen_helper_sve_compact_d }; -TRANS_FEAT(COMPACT, aa64_sve, gen_gvec_ool_arg_zpz, compact_fns[a->esz], a, 0) +TRANS_FEAT_NONSTREAMING(COMPACT, aa64_sve, gen_gvec_ool_arg_zpz, + compact_fns[a->esz], a, 0) /* Call the helper that computes the ARM LastActiveElement pseudocode * function, scaled by the element size. This includes the not found @@ -XXX,XX +XXX,XX @@ static gen_helper_gvec_3 * const bext_fns[4] = { gen_helper_sve2_bext_b, gen_helper_sve2_bext_h, gen_helper_sve2_bext_s, gen_helper_sve2_bext_d, }; -TRANS_FEAT(BEXT, aa64_sve2_bitperm, gen_gvec_ool_arg_zzz, - bext_fns[a->esz], a, 0) +TRANS_FEAT_NONSTREAMING(BEXT, aa64_sve2_bitperm, gen_gvec_ool_arg_zzz, + bext_fns[a->esz], a, 0) static gen_helper_gvec_3 * const bdep_fns[4] = { gen_helper_sve2_bdep_b, gen_helper_sve2_bdep_h, gen_helper_sve2_bdep_s, gen_helper_sve2_bdep_d, }; -TRANS_FEAT(BDEP, aa64_sve2_bitperm, gen_gvec_ool_arg_zzz, - bdep_fns[a->esz], a, 0) +TRANS_FEAT_NONSTREAMING(BDEP, aa64_sve2_bitperm, gen_gvec_ool_arg_zzz, + bdep_fns[a->esz], a, 0) static gen_helper_gvec_3 * const bgrp_fns[4] = { gen_helper_sve2_bgrp_b, gen_helper_sve2_bgrp_h, gen_helper_sve2_bgrp_s, gen_helper_sve2_bgrp_d, }; -TRANS_FEAT(BGRP, aa64_sve2_bitperm, gen_gvec_ool_arg_zzz, - bgrp_fns[a->esz], a, 0) +TRANS_FEAT_NONSTREAMING(BGRP, aa64_sve2_bitperm, gen_gvec_ool_arg_zzz, + bgrp_fns[a->esz], a, 0) static gen_helper_gvec_3 * const cadd_fns[4] = { gen_helper_sve2_cadd_b, gen_helper_sve2_cadd_h, -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-8-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 2 -- target/arm/translate-sve.c | 24 +++++++++++++++--------- 2 files changed, 15 insertions(+), 11 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 0100 0101 000- ---- 0110 1--- ---- ---- # PMULLB, PMULLT (128b result) -FAIL 0110 0100 --1- ---- 1110 01-- ---- ---- # FMMLA, BFMMLA FAIL 0110 0101 --0- ---- 0000 11-- ---- ---- # FTSMUL FAIL 0110 0101 --01 0--- 100- ---- ---- ---- # FTMAD FAIL 0110 0101 --01 1--- 001- ---- ---- ---- # FADDA 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 do_trans_pmull(DisasContext *s, arg_rrr_esz *a, bool sel) gen_helper_gvec_pmull_q, gen_helper_sve2_pmull_h, NULL, gen_helper_sve2_pmull_d, }; - if (a->esz == 0 - ? !dc_isar_feature(aa64_sve2_pmull128, s) - : !dc_isar_feature(aa64_sve, s)) { + + if (a->esz == 0) { + if (!dc_isar_feature(aa64_sve2_pmull128, s)) { + return false; + } + s->is_nonstreaming = true; + } else if (!dc_isar_feature(aa64_sve, s)) { return false; } return gen_gvec_ool_arg_zzz(s, fns[a->esz], a, sel); @@ -XXX,XX +XXX,XX @@ DO_ZPZZ_FP(FMINP, aa64_sve2, sve2_fminp_zpzz) * SVE Integer Multiply-Add (unpredicated) */ -TRANS_FEAT(FMMLA_s, aa64_sve_f32mm, gen_gvec_fpst_zzzz, gen_helper_fmmla_s, - a->rd, a->rn, a->rm, a->ra, 0, FPST_FPCR) -TRANS_FEAT(FMMLA_d, aa64_sve_f64mm, gen_gvec_fpst_zzzz, gen_helper_fmmla_d, - a->rd, a->rn, a->rm, a->ra, 0, FPST_FPCR) +TRANS_FEAT_NONSTREAMING(FMMLA_s, aa64_sve_f32mm, gen_gvec_fpst_zzzz, + gen_helper_fmmla_s, a->rd, a->rn, a->rm, a->ra, + 0, FPST_FPCR) +TRANS_FEAT_NONSTREAMING(FMMLA_d, aa64_sve_f64mm, gen_gvec_fpst_zzzz, + gen_helper_fmmla_d, a->rd, a->rn, a->rm, a->ra, + 0, FPST_FPCR) static gen_helper_gvec_4 * const sqdmlal_zzzw_fns[] = { NULL, gen_helper_sve2_sqdmlal_zzzw_h, @@ -XXX,XX +XXX,XX @@ TRANS_FEAT(BFDOT_zzzz, aa64_sve_bf16, gen_gvec_ool_arg_zzzz, TRANS_FEAT(BFDOT_zzxz, aa64_sve_bf16, gen_gvec_ool_arg_zzxz, gen_helper_gvec_bfdot_idx, a) -TRANS_FEAT(BFMMLA, aa64_sve_bf16, gen_gvec_ool_arg_zzzz, - gen_helper_gvec_bfmmla, a, 0) +TRANS_FEAT_NONSTREAMING(BFMMLA, aa64_sve_bf16, gen_gvec_ool_arg_zzzz, + gen_helper_gvec_bfmmla, a, 0) static bool do_BFMLAL_zzzw(DisasContext *s, arg_rrrr_esz *a, bool sel) { -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-9-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 3 --- target/arm/translate-sve.c | 15 +++++++++++---- 2 files changed, 11 insertions(+), 7 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 0110 0101 --0- ---- 0000 11-- ---- ---- # FTSMUL -FAIL 0110 0101 --01 0--- 100- ---- ---- ---- # FTMAD -FAIL 0110 0101 --01 1--- 001- ---- ---- ---- # FADDA FAIL 0100 0101 --0- ---- 1001 10-- ---- ---- # SMMLA, UMMLA, USMMLA FAIL 0100 0101 --1- ---- 1--- ---- ---- ---- # SVE2 string/histo/crypto instructions FAIL 1000 010- -00- ---- 10-- ---- ---- ---- # SVE2 32-bit gather NT load (vector+scalar) 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_3_ptr * const ftmad_fns[4] = { NULL, gen_helper_sve_ftmad_h, gen_helper_sve_ftmad_s, gen_helper_sve_ftmad_d, }; -TRANS_FEAT(FTMAD, aa64_sve, gen_gvec_fpst_zzz, - ftmad_fns[a->esz], a->rd, a->rn, a->rm, a->imm, - a->esz == MO_16 ? FPST_FPCR_F16 : FPST_FPCR) +TRANS_FEAT_NONSTREAMING(FTMAD, aa64_sve, gen_gvec_fpst_zzz, + ftmad_fns[a->esz], a->rd, a->rn, a->rm, a->imm, + a->esz == MO_16 ? FPST_FPCR_F16 : FPST_FPCR) /* *** SVE Floating Point Accumulating Reduction Group @@ -XXX,XX +XXX,XX @@ static bool trans_FADDA(DisasContext *s, arg_rprr_esz *a) if (a->esz == 0 || !dc_isar_feature(aa64_sve, s)) { return false; } + s->is_nonstreaming = true; if (!sve_access_check(s)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_FADDA(DisasContext *s, arg_rprr_esz *a) DO_FP3(FADD_zzz, fadd) DO_FP3(FSUB_zzz, fsub) DO_FP3(FMUL_zzz, fmul) -DO_FP3(FTSMUL, ftsmul) DO_FP3(FRECPS, recps) DO_FP3(FRSQRTS, rsqrts) #undef DO_FP3 +static gen_helper_gvec_3_ptr * const ftsmul_fns[4] = { + NULL, gen_helper_gvec_ftsmul_h, + gen_helper_gvec_ftsmul_s, gen_helper_gvec_ftsmul_d +}; +TRANS_FEAT_NONSTREAMING(FTSMUL, aa64_sve, gen_gvec_fpst_arg_zzz, + ftsmul_fns[a->esz], a, 0) + /* *** SVE Floating Point Arithmetic - Predicated Group */ -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-10-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 1 - target/arm/translate-sve.c | 12 ++++++------ 2 files changed, 6 insertions(+), 7 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 0100 0101 --0- ---- 1001 10-- ---- ---- # SMMLA, UMMLA, USMMLA FAIL 0100 0101 --1- ---- 1--- ---- ---- ---- # SVE2 string/histo/crypto instructions FAIL 1000 010- -00- ---- 10-- ---- ---- ---- # SVE2 32-bit gather NT load (vector+scalar) FAIL 1000 010- -00- ---- 111- ---- ---- ---- # SVE 32-bit gather prefetch (vector+imm) 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 @@ TRANS_FEAT(FMLALT_zzxw, aa64_sve2, do_FMLAL_zzxw, a, false, true) TRANS_FEAT(FMLSLB_zzxw, aa64_sve2, do_FMLAL_zzxw, a, true, false) TRANS_FEAT(FMLSLT_zzxw, aa64_sve2, do_FMLAL_zzxw, a, true, true) -TRANS_FEAT(SMMLA, aa64_sve_i8mm, gen_gvec_ool_arg_zzzz, - gen_helper_gvec_smmla_b, a, 0) -TRANS_FEAT(USMMLA, aa64_sve_i8mm, gen_gvec_ool_arg_zzzz, - gen_helper_gvec_usmmla_b, a, 0) -TRANS_FEAT(UMMLA, aa64_sve_i8mm, gen_gvec_ool_arg_zzzz, - gen_helper_gvec_ummla_b, a, 0) +TRANS_FEAT_NONSTREAMING(SMMLA, aa64_sve_i8mm, gen_gvec_ool_arg_zzzz, + gen_helper_gvec_smmla_b, a, 0) +TRANS_FEAT_NONSTREAMING(USMMLA, aa64_sve_i8mm, gen_gvec_ool_arg_zzzz, + gen_helper_gvec_usmmla_b, a, 0) +TRANS_FEAT_NONSTREAMING(UMMLA, aa64_sve_i8mm, gen_gvec_ool_arg_zzzz, + gen_helper_gvec_ummla_b, a, 0) TRANS_FEAT(BFDOT_zzzz, aa64_sve_bf16, gen_gvec_ool_arg_zzzz, gen_helper_gvec_bfdot, a, 0) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-11-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 1 - target/arm/translate-sve.c | 35 ++++++++++++++++++----------------- 2 files changed, 18 insertions(+), 18 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 0100 0101 --1- ---- 1--- ---- ---- ---- # SVE2 string/histo/crypto instructions FAIL 1000 010- -00- ---- 10-- ---- ---- ---- # SVE2 32-bit gather NT load (vector+scalar) FAIL 1000 010- -00- ---- 111- ---- ---- ---- # SVE 32-bit gather prefetch (vector+imm) FAIL 1000 0100 0-1- ---- 0--- ---- ---- ---- # SVE 32-bit gather prefetch (scalar+vector) 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 @@ DO_SVE2_ZZZ_NARROW(RSUBHNT, rsubhnt) static gen_helper_gvec_flags_4 * const match_fns[4] = { gen_helper_sve2_match_ppzz_b, gen_helper_sve2_match_ppzz_h, NULL, NULL }; -TRANS_FEAT(MATCH, aa64_sve2, do_ppzz_flags, a, match_fns[a->esz]) +TRANS_FEAT_NONSTREAMING(MATCH, aa64_sve2, do_ppzz_flags, a, match_fns[a->esz]) static gen_helper_gvec_flags_4 * const nmatch_fns[4] = { gen_helper_sve2_nmatch_ppzz_b, gen_helper_sve2_nmatch_ppzz_h, NULL, NULL }; -TRANS_FEAT(NMATCH, aa64_sve2, do_ppzz_flags, a, nmatch_fns[a->esz]) +TRANS_FEAT_NONSTREAMING(NMATCH, aa64_sve2, do_ppzz_flags, a, nmatch_fns[a->esz]) static gen_helper_gvec_4 * const histcnt_fns[4] = { NULL, NULL, gen_helper_sve2_histcnt_s, gen_helper_sve2_histcnt_d }; -TRANS_FEAT(HISTCNT, aa64_sve2, gen_gvec_ool_arg_zpzz, - histcnt_fns[a->esz], a, 0) +TRANS_FEAT_NONSTREAMING(HISTCNT, aa64_sve2, gen_gvec_ool_arg_zpzz, + histcnt_fns[a->esz], a, 0) -TRANS_FEAT(HISTSEG, aa64_sve2, gen_gvec_ool_arg_zzz, - a->esz == 0 ? gen_helper_sve2_histseg : NULL, a, 0) +TRANS_FEAT_NONSTREAMING(HISTSEG, aa64_sve2, gen_gvec_ool_arg_zzz, + a->esz == 0 ? gen_helper_sve2_histseg : NULL, a, 0) DO_ZPZZ_FP(FADDP, aa64_sve2, sve2_faddp_zpzz) DO_ZPZZ_FP(FMAXNMP, aa64_sve2, sve2_fmaxnmp_zpzz) @@ -XXX,XX +XXX,XX @@ TRANS_FEAT(SQRDCMLAH_zzzz, aa64_sve2, gen_gvec_ool_zzzz, TRANS_FEAT(USDOT_zzzz, aa64_sve_i8mm, gen_gvec_ool_arg_zzzz, a->esz == 2 ? gen_helper_gvec_usdot_b : NULL, a, 0) -TRANS_FEAT(AESMC, aa64_sve2_aes, gen_gvec_ool_zz, - gen_helper_crypto_aesmc, a->rd, a->rd, a->decrypt) +TRANS_FEAT_NONSTREAMING(AESMC, aa64_sve2_aes, gen_gvec_ool_zz, + gen_helper_crypto_aesmc, a->rd, a->rd, a->decrypt) -TRANS_FEAT(AESE, aa64_sve2_aes, gen_gvec_ool_arg_zzz, - gen_helper_crypto_aese, a, false) -TRANS_FEAT(AESD, aa64_sve2_aes, gen_gvec_ool_arg_zzz, - gen_helper_crypto_aese, a, true) +TRANS_FEAT_NONSTREAMING(AESE, aa64_sve2_aes, gen_gvec_ool_arg_zzz, + gen_helper_crypto_aese, a, false) +TRANS_FEAT_NONSTREAMING(AESD, aa64_sve2_aes, gen_gvec_ool_arg_zzz, + gen_helper_crypto_aese, a, true) -TRANS_FEAT(SM4E, aa64_sve2_sm4, gen_gvec_ool_arg_zzz, - gen_helper_crypto_sm4e, a, 0) -TRANS_FEAT(SM4EKEY, aa64_sve2_sm4, gen_gvec_ool_arg_zzz, - gen_helper_crypto_sm4ekey, a, 0) +TRANS_FEAT_NONSTREAMING(SM4E, aa64_sve2_sm4, gen_gvec_ool_arg_zzz, + gen_helper_crypto_sm4e, a, 0) +TRANS_FEAT_NONSTREAMING(SM4EKEY, aa64_sve2_sm4, gen_gvec_ool_arg_zzz, + gen_helper_crypto_sm4ekey, a, 0) -TRANS_FEAT(RAX1, aa64_sve2_sha3, gen_gvec_fn_arg_zzz, gen_gvec_rax1, a) +TRANS_FEAT_NONSTREAMING(RAX1, aa64_sve2_sha3, gen_gvec_fn_arg_zzz, + gen_gvec_rax1, a) TRANS_FEAT(FCVTNT_sh, aa64_sve2, gen_gvec_fpst_arg_zpz, gen_helper_sve2_fcvtnt_sh, a, 0, FPST_FPCR) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-12-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 9 --------- target/arm/translate-sve.c | 6 ++++++ 2 files changed, 6 insertions(+), 9 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 1000 010- -00- ---- 10-- ---- ---- ---- # SVE2 32-bit gather NT load (vector+scalar) FAIL 1000 010- -00- ---- 111- ---- ---- ---- # SVE 32-bit gather prefetch (vector+imm) FAIL 1000 0100 0-1- ---- 0--- ---- ---- ---- # SVE 32-bit gather prefetch (scalar+vector) -FAIL 1000 010- -01- ---- 1--- ---- ---- ---- # SVE 32-bit gather load (vector+imm) -FAIL 1000 0100 0-0- ---- 0--- ---- ---- ---- # SVE 32-bit gather load byte (scalar+vector) -FAIL 1000 0100 1--- ---- 0--- ---- ---- ---- # SVE 32-bit gather load half (scalar+vector) -FAIL 1000 0101 0--- ---- 0--- ---- ---- ---- # SVE 32-bit gather load word (scalar+vector) FAIL 1010 010- ---- ---- 011- ---- ---- ---- # SVE contiguous FF load (scalar+scalar) FAIL 1010 010- ---1 ---- 101- ---- ---- ---- # SVE contiguous NF load (scalar+imm) FAIL 1010 010- -01- ---- 000- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+scalar) FAIL 1010 010- -010 ---- 001- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+imm) FAIL 1100 010- ---- ---- ---- ---- ---- ---- # SVE 64-bit gather load/prefetch -FAIL 1110 010- -00- ---- 001- ---- ---- ---- # SVE2 64-bit scatter NT store (vector+scalar) -FAIL 1110 010- -10- ---- 001- ---- ---- ---- # SVE2 32-bit scatter NT store (vector+scalar) -FAIL 1110 010- ---- ---- 1-0- ---- ---- ---- # SVE scatter store (scalar+32-bit vector) -FAIL 1110 010- ---- ---- 101- ---- ---- ---- # SVE scatter store (misc) 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_zprz(DisasContext *s, arg_LD1_zprz *a) if (!dc_isar_feature(aa64_sve, s)) { return false; } + s->is_nonstreaming = true; if (!sve_access_check(s)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zpiz(DisasContext *s, arg_LD1_zpiz *a) if (!dc_isar_feature(aa64_sve, s)) { return false; } + s->is_nonstreaming = true; if (!sve_access_check(s)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_LDNT1_zprz(DisasContext *s, arg_LD1_zprz *a) if (!dc_isar_feature(aa64_sve2, s)) { return false; } + s->is_nonstreaming = true; if (!sve_access_check(s)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zprz(DisasContext *s, arg_ST1_zprz *a) if (!dc_isar_feature(aa64_sve, s)) { return false; } + s->is_nonstreaming = true; if (!sve_access_check(s)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zpiz(DisasContext *s, arg_ST1_zpiz *a) if (!dc_isar_feature(aa64_sve, s)) { return false; } + s->is_nonstreaming = true; if (!sve_access_check(s)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_STNT1_zprz(DisasContext *s, arg_ST1_zprz *a) if (!dc_isar_feature(aa64_sve2, s)) { return false; } + s->is_nonstreaming = true; if (!sve_access_check(s)) { return true; } -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. In this case, introduce PRF_ns (prefetch non-streaming) to handle the checks. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-13-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 3 --- target/arm/sve.decode | 10 +++++----- target/arm/translate-sve.c | 11 +++++++++++ 3 files changed, 16 insertions(+), 8 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 1000 010- -00- ---- 111- ---- ---- ---- # SVE 32-bit gather prefetch (vector+imm) -FAIL 1000 0100 0-1- ---- 0--- ---- ---- ---- # SVE 32-bit gather prefetch (scalar+vector) FAIL 1010 010- ---- ---- 011- ---- ---- ---- # SVE contiguous FF load (scalar+scalar) FAIL 1010 010- ---1 ---- 101- ---- ---- ---- # SVE contiguous NF load (scalar+imm) FAIL 1010 010- -01- ---- 000- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+scalar) FAIL 1010 010- -010 ---- 001- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+imm) -FAIL 1100 010- ---- ---- ---- ---- ---- ---- # SVE 64-bit gather load/prefetch 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 @@ LD1RO_zpri 1010010 .. 01 0.... 001 ... ..... ..... \ @rpri_load_msz nreg=0 # SVE 32-bit gather prefetch (scalar plus 32-bit scaled offsets) -PRF 1000010 00 -1 ----- 0-- --- ----- 0 ---- +PRF_ns 1000010 00 -1 ----- 0-- --- ----- 0 ---- # SVE 32-bit gather prefetch (vector plus immediate) -PRF 1000010 -- 00 ----- 111 --- ----- 0 ---- +PRF_ns 1000010 -- 00 ----- 111 --- ----- 0 ---- # SVE contiguous prefetch (scalar plus immediate) PRF 1000010 11 1- ----- 0-- --- ----- 0 ---- @@ -XXX,XX +XXX,XX @@ LD1_zpiz 1100010 .. 01 ..... 1.. ... ..... ..... \ @rpri_g_load esz=3 # SVE 64-bit gather prefetch (scalar plus 64-bit scaled offsets) -PRF 1100010 00 11 ----- 1-- --- ----- 0 ---- +PRF_ns 1100010 00 11 ----- 1-- --- ----- 0 ---- # SVE 64-bit gather prefetch (scalar plus unpacked 32-bit scaled offsets) -PRF 1100010 00 -1 ----- 0-- --- ----- 0 ---- +PRF_ns 1100010 00 -1 ----- 0-- --- ----- 0 ---- # SVE 64-bit gather prefetch (vector plus immediate) -PRF 1100010 -- 00 ----- 111 --- ----- 0 ---- +PRF_ns 1100010 -- 00 ----- 111 --- ----- 0 ---- ### SVE Memory Store Group 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_PRF_rr(DisasContext *s, arg_PRF_rr *a) return true; } +static bool trans_PRF_ns(DisasContext *s, arg_PRF_ns *a) +{ + if (!dc_isar_feature(aa64_sve, s)) { + return false; + } + /* Prefetch is a nop within QEMU. */ + s->is_nonstreaming = true; + (void)sve_access_check(s); + return true; +} + /* * Move Prefix * -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-14-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 2 -- target/arm/translate-sve.c | 2 ++ 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 1010 010- ---- ---- 011- ---- ---- ---- # SVE contiguous FF load (scalar+scalar) -FAIL 1010 010- ---1 ---- 101- ---- ---- ---- # SVE contiguous NF load (scalar+imm) FAIL 1010 010- -01- ---- 000- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+scalar) FAIL 1010 010- -010 ---- 001- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+imm) 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_LDFF1_zprr(DisasContext *s, arg_rprr_load *a) if (!dc_isar_feature(aa64_sve, s)) { return false; } + s->is_nonstreaming = true; 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)); @@ -XXX,XX +XXX,XX @@ static bool trans_LDNF1_zpri(DisasContext *s, arg_rpri_load *a) if (!dc_isar_feature(aa64_sve, s)) { return false; } + s->is_nonstreaming = true; if (sve_access_check(s)) { int vsz = vec_full_reg_size(s); int elements = vsz >> dtype_esz[a->dtype]; -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-15-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 3 --- target/arm/translate-sve.c | 2 ++ 2 files changed, 2 insertions(+), 3 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --0- ---- ---- ---- ---- ---- # Load/store FP register (unscaled imm) # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) - -FAIL 1010 010- -01- ---- 000- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+scalar) -FAIL 1010 010- -010 ---- 001- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+imm) 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_LD1RO_zprr(DisasContext *s, arg_rprr_load *a) if (a->rm == 31) { return false; } + s->is_nonstreaming = true; 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)); @@ -XXX,XX +XXX,XX @@ static bool trans_LD1RO_zpri(DisasContext *s, arg_rpri_load *a) if (!dc_isar_feature(aa64_sve_f64mm, s)) { return false; } + s->is_nonstreaming = true; if (sve_access_check(s)) { TCGv_i64 addr = new_tmp_a64(s); tcg_gen_addi_i64(addr, cpu_reg_sp(s, a->rn), a->imm * 32); -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> These functions will be used to verify that the cpu is in the correct state for a given instruction. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-16-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.h | 21 +++++++++++++++++++++ target/arm/translate-a64.c | 34 ++++++++++++++++++++++++++++++++++ 2 files changed, 55 insertions(+) diff --git a/target/arm/translate-a64.h b/target/arm/translate-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.h +++ b/target/arm/translate-a64.h @@ -XXX,XX +XXX,XX @@ void write_fp_dreg(DisasContext *s, int reg, TCGv_i64 v); bool logic_imm_decode_wmask(uint64_t *result, unsigned int immn, unsigned int imms, unsigned int immr); bool sve_access_check(DisasContext *s); +bool sme_enabled_check(DisasContext *s); +bool sme_enabled_check_with_svcr(DisasContext *s, unsigned); + +/* This function corresponds to CheckStreamingSVEEnabled. */ +static inline bool sme_sm_enabled_check(DisasContext *s) +{ + return sme_enabled_check_with_svcr(s, R_SVCR_SM_MASK); +} + +/* This function corresponds to CheckSMEAndZAEnabled. */ +static inline bool sme_za_enabled_check(DisasContext *s) +{ + return sme_enabled_check_with_svcr(s, R_SVCR_ZA_MASK); +} + +/* Note that this function corresponds to CheckStreamingSVEAndZAEnabled. */ +static inline bool sme_smza_enabled_check(DisasContext *s) +{ + return sme_enabled_check_with_svcr(s, R_SVCR_SM_MASK | R_SVCR_ZA_MASK); +} + TCGv_i64 clean_data_tbi(DisasContext *s, TCGv_i64 addr); TCGv_i64 gen_mte_check1(DisasContext *s, TCGv_i64 addr, bool is_write, bool tag_checked, int log2_size); 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 bool sme_access_check(DisasContext *s) return true; } +/* This function corresponds to CheckSMEEnabled. */ +bool sme_enabled_check(DisasContext *s) +{ + /* + * Note that unlike sve_excp_el, we have not constrained sme_excp_el + * to be zero when fp_excp_el has priority. This is because we need + * sme_excp_el by itself for cpregs access checks. + */ + if (!s->fp_excp_el || s->sme_excp_el < s->fp_excp_el) { + s->fp_access_checked = true; + return sme_access_check(s); + } + return fp_access_check_only(s); +} + +/* Common subroutine for CheckSMEAnd*Enabled. */ +bool sme_enabled_check_with_svcr(DisasContext *s, unsigned req) +{ + if (!sme_enabled_check(s)) { + return false; + } + if (FIELD_EX64(req, SVCR, SM) && !s->pstate_sm) { + gen_exception_insn(s, s->pc_curr, EXCP_UDEF, + syn_smetrap(SME_ET_NotStreaming, false)); + return false; + } + if (FIELD_EX64(req, SVCR, ZA) && !s->pstate_za) { + gen_exception_insn(s, s->pc_curr, EXCP_UDEF, + syn_smetrap(SME_ET_InactiveZA, false)); + return false; + } + return true; +} + /* * This utility function is for doing register extension with an * optional shift. You will likely want to pass a temporary for the -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> The pseudocode for CheckSVEEnabled gains a check for Streaming SVE mode, and for SME present but SVE absent. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-17-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 22 ++++++++++++++++------ 1 file changed, 16 insertions(+), 6 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static bool fp_access_check(DisasContext *s) return true; } -/* Check that SVE access is enabled. If it is, return true. +/* + * Check that SVE access is enabled. If it is, return true. * If not, emit code to generate an appropriate exception and return false. + * This function corresponds to CheckSVEEnabled(). */ bool sve_access_check(DisasContext *s) { - if (s->sve_excp_el) { - assert(!s->sve_access_checked); - s->sve_access_checked = true; - + if (s->pstate_sm || !dc_isar_feature(aa64_sve, s)) { + assert(dc_isar_feature(aa64_sme, s)); + if (!sme_sm_enabled_check(s)) { + goto fail_exit; + } + } else if (s->sve_excp_el) { gen_exception_insn_el(s, s->pc_curr, EXCP_UDEF, syn_sve_access_trap(), s->sve_excp_el); - return false; + goto fail_exit; } s->sve_access_checked = true; return fp_access_check(s); + + fail_exit: + /* Assert that we only raise one exception per instruction. */ + assert(!s->sve_access_checked); + s->sve_access_checked = true; + return false; } /* -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> These SME instructions are nominally within the SVE decode space, so we add them to sve.decode and translate-sve.c. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-18-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.h | 12 ++++++++++++ target/arm/sve.decode | 5 ++++- target/arm/translate-sve.c | 38 ++++++++++++++++++++++++++++++++++++++ 3 files changed, 54 insertions(+), 1 deletion(-) diff --git a/target/arm/translate-a64.h b/target/arm/translate-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.h +++ b/target/arm/translate-a64.h @@ -XXX,XX +XXX,XX @@ static inline int vec_full_reg_size(DisasContext *s) return s->vl; } +/* Return the byte size of the vector register, SVL / 8. */ +static inline int streaming_vec_reg_size(DisasContext *s) +{ + return s->svl; +} + /* * Return the offset info CPUARMState of the predicate vector register Pn. * Note for this purpose, FFR is P16. @@ -XXX,XX +XXX,XX @@ static inline int pred_full_reg_size(DisasContext *s) return s->vl >> 3; } +/* Return the byte size of the predicate register, SVL / 64. */ +static inline int streaming_pred_reg_size(DisasContext *s) +{ + return s->svl >> 3; +} + /* * Round up the size of a register to a size allowed by * the tcg vector infrastructure. Any operation which uses this 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 @@ INDEX_ri 00000100 esz:2 1 imm:s5 010001 rn:5 rd:5 # SVE index generation (register start, register increment) INDEX_rr 00000100 .. 1 ..... 010011 ..... ..... @rd_rn_rm -### SVE Stack Allocation Group +### SVE / Streaming SVE Stack Allocation Group # SVE stack frame adjustment ADDVL 00000100 001 ..... 01010 ...... ..... @rd_rn_i6 +ADDSVL 00000100 001 ..... 01011 ...... ..... @rd_rn_i6 ADDPL 00000100 011 ..... 01010 ...... ..... @rd_rn_i6 +ADDSPL 00000100 011 ..... 01011 ...... ..... @rd_rn_i6 # SVE stack frame size RDVL 00000100 101 11111 01010 imm:s6 rd:5 +RDSVL 00000100 101 11111 01011 imm:s6 rd:5 ### SVE Bitwise Shift - Unpredicated Group 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_ADDVL(DisasContext *s, arg_ADDVL *a) return true; } +static bool trans_ADDSVL(DisasContext *s, arg_ADDSVL *a) +{ + if (!dc_isar_feature(aa64_sme, s)) { + return false; + } + if (sme_enabled_check(s)) { + TCGv_i64 rd = cpu_reg_sp(s, a->rd); + TCGv_i64 rn = cpu_reg_sp(s, a->rn); + tcg_gen_addi_i64(rd, rn, a->imm * streaming_vec_reg_size(s)); + } + return true; +} + static bool trans_ADDPL(DisasContext *s, arg_ADDPL *a) { if (!dc_isar_feature(aa64_sve, s)) { @@ -XXX,XX +XXX,XX @@ static bool trans_ADDPL(DisasContext *s, arg_ADDPL *a) return true; } +static bool trans_ADDSPL(DisasContext *s, arg_ADDSPL *a) +{ + if (!dc_isar_feature(aa64_sme, s)) { + return false; + } + if (sme_enabled_check(s)) { + TCGv_i64 rd = cpu_reg_sp(s, a->rd); + TCGv_i64 rn = cpu_reg_sp(s, a->rn); + tcg_gen_addi_i64(rd, rn, a->imm * streaming_pred_reg_size(s)); + } + return true; +} + static bool trans_RDVL(DisasContext *s, arg_RDVL *a) { if (!dc_isar_feature(aa64_sve, s)) { @@ -XXX,XX +XXX,XX @@ static bool trans_RDVL(DisasContext *s, arg_RDVL *a) return true; } +static bool trans_RDSVL(DisasContext *s, arg_RDSVL *a) +{ + if (!dc_isar_feature(aa64_sme, s)) { + return false; + } + if (sme_enabled_check(s)) { + TCGv_i64 reg = cpu_reg(s, a->rd); + tcg_gen_movi_i64(reg, a->imm * streaming_vec_reg_size(s)); + } + return true; +} + /* *** SVE Compute Vector Address Group */ -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-19-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 2 ++ target/arm/sme.decode | 4 ++++ target/arm/sme_helper.c | 25 +++++++++++++++++++++++++ target/arm/translate-sme.c | 13 +++++++++++++ 4 files changed, 44 insertions(+) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_2(set_pstate_sm, TCG_CALL_NO_RWG, void, env, i32) DEF_HELPER_FLAGS_2(set_pstate_za, TCG_CALL_NO_RWG, void, env, i32) + +DEF_HELPER_FLAGS_3(sme_zero, TCG_CALL_NO_RWG, void, env, i32, i32) diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ # # This file is processed by scripts/decodetree.py # + +### SME Misc + +ZERO 11000000 00 001 00000000000 imm:8 diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ void helper_set_pstate_za(CPUARMState *env, uint32_t i) memset(env->zarray, 0, sizeof(env->zarray)); } } + +void helper_sme_zero(CPUARMState *env, uint32_t imm, uint32_t svl) +{ + uint32_t i; + + /* + * Special case clearing the entire ZA space. + * This falls into the CONSTRAINED UNPREDICTABLE zeroing of any + * parts of the ZA storage outside of SVL. + */ + if (imm == 0xff) { + memset(env->zarray, 0, sizeof(env->zarray)); + return; + } + + /* + * Recall that ZAnH.D[m] is spread across ZA[n+8*m], + * so each row is discontiguous within ZA[]. + */ + for (i = 0; i < svl; i++) { + if (imm & (1 << (i % 8))) { + memset(&env->zarray[i], 0, svl); + } + } +} diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ */ #include "decode-sme.c.inc" + + +static bool trans_ZERO(DisasContext *s, arg_ZERO *a) +{ + if (!dc_isar_feature(aa64_sme, s)) { + return false; + } + if (sme_za_enabled_check(s)) { + gen_helper_sme_zero(cpu_env, tcg_constant_i32(a->imm), + tcg_constant_i32(streaming_vec_reg_size(s))); + } + return true; +} -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> We can reuse the SVE functions for implementing moves to/from horizontal tile slices, but we need new ones for moves to/from vertical tile slices. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-20-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 12 +++ target/arm/helper-sve.h | 2 + target/arm/translate-a64.h | 8 ++ target/arm/translate.h | 5 ++ target/arm/sme.decode | 15 ++++ target/arm/sme_helper.c | 151 ++++++++++++++++++++++++++++++++++++- target/arm/sve_helper.c | 12 +++ target/arm/translate-sme.c | 127 +++++++++++++++++++++++++++++++ 8 files changed, 331 insertions(+), 1 deletion(-) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_2(set_pstate_sm, TCG_CALL_NO_RWG, void, env, i32) DEF_HELPER_FLAGS_2(set_pstate_za, TCG_CALL_NO_RWG, void, env, i32) DEF_HELPER_FLAGS_3(sme_zero, TCG_CALL_NO_RWG, void, env, i32, i32) + +/* Move to/from vertical array slices, i.e. columns, so 'c'. */ +DEF_HELPER_FLAGS_4(sme_mova_cz_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_zc_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_cz_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_zc_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_cz_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_zc_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_cz_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_zc_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_cz_q, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_zc_q, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) 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_5(sve_sel_zpzz_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sve_sel_zpzz_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(sve_sel_zpzz_q, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sve2_addp_zpzz_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) diff --git a/target/arm/translate-a64.h b/target/arm/translate-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.h +++ b/target/arm/translate-a64.h @@ -XXX,XX +XXX,XX @@ static inline int pred_gvec_reg_size(DisasContext *s) return size_for_gvec(pred_full_reg_size(s)); } +/* Return a newly allocated pointer to the predicate register. */ +static inline TCGv_ptr pred_full_reg_ptr(DisasContext *s, int regno) +{ + TCGv_ptr ret = tcg_temp_new_ptr(); + tcg_gen_addi_ptr(ret, cpu_env, pred_full_reg_offset(s, regno)); + return ret; +} + bool disas_sve(DisasContext *, uint32_t); bool disas_sme(DisasContext *, uint32_t); 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 @@ static inline int plus_2(DisasContext *s, int x) return x + 2; } +static inline int plus_12(DisasContext *s, int x) +{ + return x + 12; +} + static inline int times_2(DisasContext *s, int x) { return x * 2; diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ ### SME Misc ZERO 11000000 00 001 00000000000 imm:8 + +### SME Move into/from Array + +%mova_rs 13:2 !function=plus_12 +&mova esz rs pg zr za_imm v:bool to_vec:bool + +MOVA 11000000 esz:2 00000 0 v:1 .. pg:3 zr:5 0 za_imm:4 \ + &mova to_vec=0 rs=%mova_rs +MOVA 11000000 11 00000 1 v:1 .. pg:3 zr:5 0 za_imm:4 \ + &mova to_vec=0 rs=%mova_rs esz=4 + +MOVA 11000000 esz:2 00001 0 v:1 .. pg:3 0 za_imm:4 zr:5 \ + &mova to_vec=1 rs=%mova_rs +MOVA 11000000 11 00001 1 v:1 .. pg:3 0 za_imm:4 zr:5 \ + &mova to_vec=1 rs=%mova_rs esz=4 diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" -#include "internals.h" +#include "tcg/tcg-gvec-desc.h" #include "exec/helper-proto.h" +#include "qemu/int128.h" +#include "vec_internal.h" /* ResetSVEState */ void arm_reset_sve_state(CPUARMState *env) @@ -XXX,XX +XXX,XX @@ void helper_sme_zero(CPUARMState *env, uint32_t imm, uint32_t svl) } } } + + +/* + * When considering the ZA storage as an array of elements of + * type T, the index within that array of the Nth element of + * a vertical slice of a tile can be calculated like this, + * regardless of the size of type T. This is because the tiles + * are interleaved, so if type T is size N bytes then row 1 of + * the tile is N rows away from row 0. The division by N to + * convert a byte offset into an array index and the multiplication + * by N to convert from vslice-index-within-the-tile to + * the index within the ZA storage cancel out. + */ +#define tile_vslice_index(i) ((i) * sizeof(ARMVectorReg)) + +/* + * When doing byte arithmetic on the ZA storage, the element + * byteoff bytes away in a tile vertical slice is always this + * many bytes away in the ZA storage, regardless of the + * size of the tile element, assuming that byteoff is a multiple + * of the element size. Again this is because of the interleaving + * of the tiles. For instance if we have 1 byte per element then + * each row of the ZA storage has one byte of the vslice data, + * and (counting from 0) byte 8 goes in row 8 of the storage + * at offset (8 * row-size-in-bytes). + * If we have 8 bytes per element then each row of the ZA storage + * has 8 bytes of the data, but there are 8 interleaved tiles and + * so byte 8 of the data goes into row 1 of the tile, + * which is again row 8 of the storage, so the offset is still + * (8 * row-size-in-bytes). Similarly for other element sizes. + */ +#define tile_vslice_offset(byteoff) ((byteoff) * sizeof(ARMVectorReg)) + + +/* + * Move Zreg vector to ZArray column. + */ +#define DO_MOVA_C(NAME, TYPE, H) \ +void HELPER(NAME)(void *za, void *vn, void *vg, uint32_t desc) \ +{ \ + int i, oprsz = simd_oprsz(desc); \ + for (i = 0; i < oprsz; ) { \ + uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ + do { \ + if (pg & 1) { \ + *(TYPE *)(za + tile_vslice_offset(i)) = *(TYPE *)(vn + H(i)); \ + } \ + i += sizeof(TYPE); \ + pg >>= sizeof(TYPE); \ + } while (i & 15); \ + } \ +} + +DO_MOVA_C(sme_mova_cz_b, uint8_t, H1) +DO_MOVA_C(sme_mova_cz_h, uint16_t, H1_2) +DO_MOVA_C(sme_mova_cz_s, uint32_t, H1_4) + +void HELPER(sme_mova_cz_d)(void *za, void *vn, void *vg, uint32_t desc) +{ + int i, oprsz = simd_oprsz(desc) / 8; + uint8_t *pg = vg; + uint64_t *n = vn; + uint64_t *a = za; + + for (i = 0; i < oprsz; i++) { + if (pg[H1(i)] & 1) { + a[tile_vslice_index(i)] = n[i]; + } + } +} + +void HELPER(sme_mova_cz_q)(void *za, void *vn, void *vg, uint32_t desc) +{ + int i, oprsz = simd_oprsz(desc) / 16; + uint16_t *pg = vg; + Int128 *n = vn; + Int128 *a = za; + + /* + * Int128 is used here simply to copy 16 bytes, and to simplify + * the address arithmetic. + */ + for (i = 0; i < oprsz; i++) { + if (pg[H2(i)] & 1) { + a[tile_vslice_index(i)] = n[i]; + } + } +} + +#undef DO_MOVA_C + +/* + * Move ZArray column to Zreg vector. + */ +#define DO_MOVA_Z(NAME, TYPE, H) \ +void HELPER(NAME)(void *vd, void *za, void *vg, uint32_t desc) \ +{ \ + int i, oprsz = simd_oprsz(desc); \ + for (i = 0; i < oprsz; ) { \ + uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ + do { \ + if (pg & 1) { \ + *(TYPE *)(vd + H(i)) = *(TYPE *)(za + tile_vslice_offset(i)); \ + } \ + i += sizeof(TYPE); \ + pg >>= sizeof(TYPE); \ + } while (i & 15); \ + } \ +} + +DO_MOVA_Z(sme_mova_zc_b, uint8_t, H1) +DO_MOVA_Z(sme_mova_zc_h, uint16_t, H1_2) +DO_MOVA_Z(sme_mova_zc_s, uint32_t, H1_4) + +void HELPER(sme_mova_zc_d)(void *vd, void *za, void *vg, uint32_t desc) +{ + int i, oprsz = simd_oprsz(desc) / 8; + uint8_t *pg = vg; + uint64_t *d = vd; + uint64_t *a = za; + + for (i = 0; i < oprsz; i++) { + if (pg[H1(i)] & 1) { + d[i] = a[tile_vslice_index(i)]; + } + } +} + +void HELPER(sme_mova_zc_q)(void *vd, void *za, void *vg, uint32_t desc) +{ + int i, oprsz = simd_oprsz(desc) / 16; + uint16_t *pg = vg; + Int128 *d = vd; + Int128 *a = za; + + /* + * Int128 is used here simply to copy 16 bytes, and to simplify + * the address arithmetic. + */ + for (i = 0; i < oprsz; i++, za += sizeof(ARMVectorReg)) { + if (pg[H2(i)] & 1) { + d[i] = a[tile_vslice_index(i)]; + } + } +} + +#undef DO_MOVA_Z 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_sel_zpzz_d)(void *vd, void *vn, void *vm, } } +void HELPER(sve_sel_zpzz_q)(void *vd, void *vn, void *vm, + void *vg, uint32_t desc) +{ + intptr_t i, opr_sz = simd_oprsz(desc) / 16; + Int128 *d = vd, *n = vn, *m = vm; + uint16_t *pg = vg; + + for (i = 0; i < opr_sz; i += 1) { + d[i] = (pg[H2(i)] & 1 ? n : m)[i]; + } +} + /* Two operand comparison controlled by a predicate. * ??? It is very tempting to want to be able to expand this inline * with x86 instructions, e.g. diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ #include "decode-sme.c.inc" +/* + * Resolve tile.size[index] to a host pointer, where tile and index + * are always decoded together, dependent on the element size. + */ +static TCGv_ptr get_tile_rowcol(DisasContext *s, int esz, int rs, + int tile_index, bool vertical) +{ + int tile = tile_index >> (4 - esz); + int index = esz == MO_128 ? 0 : extract32(tile_index, 0, 4 - esz); + int pos, len, offset; + TCGv_i32 tmp; + TCGv_ptr addr; + + /* Compute the final index, which is Rs+imm. */ + tmp = tcg_temp_new_i32(); + tcg_gen_trunc_tl_i32(tmp, cpu_reg(s, rs)); + tcg_gen_addi_i32(tmp, tmp, index); + + /* Prepare a power-of-two modulo via extraction of @len bits. */ + len = ctz32(streaming_vec_reg_size(s)) - esz; + + if (vertical) { + /* + * Compute the byte offset of the index within the tile: + * (index % (svl / size)) * size + * = (index % (svl >> esz)) << esz + * Perform the power-of-two modulo via extraction of the low @len bits. + * Perform the multiply by shifting left by @pos bits. + * Perform these operations simultaneously via deposit into zero. + */ + pos = esz; + tcg_gen_deposit_z_i32(tmp, tmp, pos, len); + + /* + * For big-endian, adjust the indexed column byte offset within + * the uint64_t host words that make up env->zarray[]. + */ + if (HOST_BIG_ENDIAN && esz < MO_64) { + tcg_gen_xori_i32(tmp, tmp, 8 - (1 << esz)); + } + } else { + /* + * Compute the byte offset of the index within the tile: + * (index % (svl / size)) * (size * sizeof(row)) + * = (index % (svl >> esz)) << (esz + log2(sizeof(row))) + */ + pos = esz + ctz32(sizeof(ARMVectorReg)); + tcg_gen_deposit_z_i32(tmp, tmp, pos, len); + + /* Row slices are always aligned and need no endian adjustment. */ + } + + /* The tile byte offset within env->zarray is the row. */ + offset = tile * sizeof(ARMVectorReg); + + /* Include the byte offset of zarray to make this relative to env. */ + offset += offsetof(CPUARMState, zarray); + tcg_gen_addi_i32(tmp, tmp, offset); + + /* Add the byte offset to env to produce the final pointer. */ + addr = tcg_temp_new_ptr(); + tcg_gen_ext_i32_ptr(addr, tmp); + tcg_temp_free_i32(tmp); + tcg_gen_add_ptr(addr, addr, cpu_env); + + return addr; +} + static bool trans_ZERO(DisasContext *s, arg_ZERO *a) { if (!dc_isar_feature(aa64_sme, s)) { @@ -XXX,XX +XXX,XX @@ static bool trans_ZERO(DisasContext *s, arg_ZERO *a) } return true; } + +static bool trans_MOVA(DisasContext *s, arg_MOVA *a) +{ + static gen_helper_gvec_4 * const h_fns[5] = { + gen_helper_sve_sel_zpzz_b, gen_helper_sve_sel_zpzz_h, + gen_helper_sve_sel_zpzz_s, gen_helper_sve_sel_zpzz_d, + gen_helper_sve_sel_zpzz_q + }; + static gen_helper_gvec_3 * const cz_fns[5] = { + gen_helper_sme_mova_cz_b, gen_helper_sme_mova_cz_h, + gen_helper_sme_mova_cz_s, gen_helper_sme_mova_cz_d, + gen_helper_sme_mova_cz_q, + }; + static gen_helper_gvec_3 * const zc_fns[5] = { + gen_helper_sme_mova_zc_b, gen_helper_sme_mova_zc_h, + gen_helper_sme_mova_zc_s, gen_helper_sme_mova_zc_d, + gen_helper_sme_mova_zc_q, + }; + + TCGv_ptr t_za, t_zr, t_pg; + TCGv_i32 t_desc; + int svl; + + if (!dc_isar_feature(aa64_sme, s)) { + return false; + } + if (!sme_smza_enabled_check(s)) { + return true; + } + + t_za = get_tile_rowcol(s, a->esz, a->rs, a->za_imm, a->v); + t_zr = vec_full_reg_ptr(s, a->zr); + t_pg = pred_full_reg_ptr(s, a->pg); + + svl = streaming_vec_reg_size(s); + t_desc = tcg_constant_i32(simd_desc(svl, svl, 0)); + + if (a->v) { + /* Vertical slice -- use sme mova helpers. */ + if (a->to_vec) { + zc_fns[a->esz](t_zr, t_za, t_pg, t_desc); + } else { + cz_fns[a->esz](t_za, t_zr, t_pg, t_desc); + } + } else { + /* Horizontal slice -- reuse sve sel helpers. */ + if (a->to_vec) { + h_fns[a->esz](t_zr, t_za, t_zr, t_pg, t_desc); + } else { + h_fns[a->esz](t_za, t_zr, t_za, t_pg, t_desc); + } + } + + tcg_temp_free_ptr(t_za); + tcg_temp_free_ptr(t_zr); + tcg_temp_free_ptr(t_pg); + + return true; +} -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> We cannot reuse the SVE functions for LD[1-4] and ST[1-4], because those functions accept only a Zreg register number. For SME, we want to pass a pointer into ZA storage. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-21-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 82 +++++ target/arm/sme.decode | 9 + target/arm/sme_helper.c | 595 +++++++++++++++++++++++++++++++++++++ target/arm/translate-sme.c | 70 +++++ 4 files changed, 756 insertions(+) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sme_mova_cz_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sme_mova_zc_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sme_mova_cz_q, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sme_mova_zc_q, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_5(sme_ld1b_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1b_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1b_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1b_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_ld1h_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1h_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1h_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1h_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1h_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1h_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1h_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1h_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_ld1s_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1s_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1s_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1s_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1s_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1s_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1s_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1s_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_ld1d_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1d_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1d_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1d_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1d_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1d_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1d_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1d_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_ld1q_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1q_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1q_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1q_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1q_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1q_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1q_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1q_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_st1b_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1b_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1b_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1b_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_st1h_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1h_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1h_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1h_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1h_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1h_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1h_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1h_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_st1s_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1s_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1s_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1s_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1s_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1s_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1s_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1s_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_st1d_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1d_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1d_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1d_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1d_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1d_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1d_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1d_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_st1q_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1q_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1q_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1q_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1q_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1q_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1q_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1q_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ MOVA 11000000 esz:2 00001 0 v:1 .. pg:3 0 za_imm:4 zr:5 \ &mova to_vec=1 rs=%mova_rs MOVA 11000000 11 00001 1 v:1 .. pg:3 0 za_imm:4 zr:5 \ &mova to_vec=1 rs=%mova_rs esz=4 + +### SME Memory + +&ldst esz rs pg rn rm za_imm v:bool st:bool + +LDST1 1110000 0 esz:2 st:1 rm:5 v:1 .. pg:3 rn:5 0 za_imm:4 \ + &ldst rs=%mova_rs +LDST1 1110000 111 st:1 rm:5 v:1 .. pg:3 rn:5 0 za_imm:4 \ + &ldst esz=4 rs=%mova_rs diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" +#include "internals.h" #include "tcg/tcg-gvec-desc.h" #include "exec/helper-proto.h" +#include "exec/cpu_ldst.h" +#include "exec/exec-all.h" #include "qemu/int128.h" #include "vec_internal.h" +#include "sve_ldst_internal.h" /* ResetSVEState */ void arm_reset_sve_state(CPUARMState *env) @@ -XXX,XX +XXX,XX @@ void HELPER(sme_mova_zc_q)(void *vd, void *za, void *vg, uint32_t desc) } #undef DO_MOVA_Z + +/* + * Clear elements in a tile slice comprising len bytes. + */ + +typedef void ClearFn(void *ptr, size_t off, size_t len); + +static void clear_horizontal(void *ptr, size_t off, size_t len) +{ + memset(ptr + off, 0, len); +} + +static void clear_vertical_b(void *vptr, size_t off, size_t len) +{ + for (size_t i = 0; i < len; ++i) { + *(uint8_t *)(vptr + tile_vslice_offset(i + off)) = 0; + } +} + +static void clear_vertical_h(void *vptr, size_t off, size_t len) +{ + for (size_t i = 0; i < len; i += 2) { + *(uint16_t *)(vptr + tile_vslice_offset(i + off)) = 0; + } +} + +static void clear_vertical_s(void *vptr, size_t off, size_t len) +{ + for (size_t i = 0; i < len; i += 4) { + *(uint32_t *)(vptr + tile_vslice_offset(i + off)) = 0; + } +} + +static void clear_vertical_d(void *vptr, size_t off, size_t len) +{ + for (size_t i = 0; i < len; i += 8) { + *(uint64_t *)(vptr + tile_vslice_offset(i + off)) = 0; + } +} + +static void clear_vertical_q(void *vptr, size_t off, size_t len) +{ + for (size_t i = 0; i < len; i += 16) { + memset(vptr + tile_vslice_offset(i + off), 0, 16); + } +} + +/* + * Copy elements from an array into a tile slice comprising len bytes. + */ + +typedef void CopyFn(void *dst, const void *src, size_t len); + +static void copy_horizontal(void *dst, const void *src, size_t len) +{ + memcpy(dst, src, len); +} + +static void copy_vertical_b(void *vdst, const void *vsrc, size_t len) +{ + const uint8_t *src = vsrc; + uint8_t *dst = vdst; + size_t i; + + for (i = 0; i < len; ++i) { + dst[tile_vslice_index(i)] = src[i]; + } +} + +static void copy_vertical_h(void *vdst, const void *vsrc, size_t len) +{ + const uint16_t *src = vsrc; + uint16_t *dst = vdst; + size_t i; + + for (i = 0; i < len / 2; ++i) { + dst[tile_vslice_index(i)] = src[i]; + } +} + +static void copy_vertical_s(void *vdst, const void *vsrc, size_t len) +{ + const uint32_t *src = vsrc; + uint32_t *dst = vdst; + size_t i; + + for (i = 0; i < len / 4; ++i) { + dst[tile_vslice_index(i)] = src[i]; + } +} + +static void copy_vertical_d(void *vdst, const void *vsrc, size_t len) +{ + const uint64_t *src = vsrc; + uint64_t *dst = vdst; + size_t i; + + for (i = 0; i < len / 8; ++i) { + dst[tile_vslice_index(i)] = src[i]; + } +} + +static void copy_vertical_q(void *vdst, const void *vsrc, size_t len) +{ + for (size_t i = 0; i < len; i += 16) { + memcpy(vdst + tile_vslice_offset(i), vsrc + i, 16); + } +} + +/* + * Host and TLB primitives for vertical tile slice addressing. + */ + +#define DO_LD(NAME, TYPE, HOST, TLB) \ +static inline void sme_##NAME##_v_host(void *za, intptr_t off, void *host) \ +{ \ + TYPE val = HOST(host); \ + *(TYPE *)(za + tile_vslice_offset(off)) = val; \ +} \ +static inline void sme_##NAME##_v_tlb(CPUARMState *env, void *za, \ + intptr_t off, target_ulong addr, uintptr_t ra) \ +{ \ + TYPE val = TLB(env, useronly_clean_ptr(addr), ra); \ + *(TYPE *)(za + tile_vslice_offset(off)) = val; \ +} + +#define DO_ST(NAME, TYPE, HOST, TLB) \ +static inline void sme_##NAME##_v_host(void *za, intptr_t off, void *host) \ +{ \ + TYPE val = *(TYPE *)(za + tile_vslice_offset(off)); \ + HOST(host, val); \ +} \ +static inline void sme_##NAME##_v_tlb(CPUARMState *env, void *za, \ + intptr_t off, target_ulong addr, uintptr_t ra) \ +{ \ + TYPE val = *(TYPE *)(za + tile_vslice_offset(off)); \ + TLB(env, useronly_clean_ptr(addr), val, ra); \ +} + +/* + * The ARMVectorReg elements are stored in host-endian 64-bit units. + * For 128-bit quantities, the sequence defined by the Elem[] pseudocode + * corresponds to storing the two 64-bit pieces in little-endian order. + */ +#define DO_LDQ(HNAME, VNAME, BE, HOST, TLB) \ +static inline void HNAME##_host(void *za, intptr_t off, void *host) \ +{ \ + uint64_t val0 = HOST(host), val1 = HOST(host + 8); \ + uint64_t *ptr = za + off; \ + ptr[0] = BE ? val1 : val0, ptr[1] = BE ? val0 : val1; \ +} \ +static inline void VNAME##_v_host(void *za, intptr_t off, void *host) \ +{ \ + HNAME##_host(za, tile_vslice_offset(off), host); \ +} \ +static inline void HNAME##_tlb(CPUARMState *env, void *za, intptr_t off, \ + target_ulong addr, uintptr_t ra) \ +{ \ + uint64_t val0 = TLB(env, useronly_clean_ptr(addr), ra); \ + uint64_t val1 = TLB(env, useronly_clean_ptr(addr + 8), ra); \ + uint64_t *ptr = za + off; \ + ptr[0] = BE ? val1 : val0, ptr[1] = BE ? val0 : val1; \ +} \ +static inline void VNAME##_v_tlb(CPUARMState *env, void *za, intptr_t off, \ + target_ulong addr, uintptr_t ra) \ +{ \ + HNAME##_tlb(env, za, tile_vslice_offset(off), addr, ra); \ +} + +#define DO_STQ(HNAME, VNAME, BE, HOST, TLB) \ +static inline void HNAME##_host(void *za, intptr_t off, void *host) \ +{ \ + uint64_t *ptr = za + off; \ + HOST(host, ptr[BE]); \ + HOST(host + 1, ptr[!BE]); \ +} \ +static inline void VNAME##_v_host(void *za, intptr_t off, void *host) \ +{ \ + HNAME##_host(za, tile_vslice_offset(off), host); \ +} \ +static inline void HNAME##_tlb(CPUARMState *env, void *za, intptr_t off, \ + target_ulong addr, uintptr_t ra) \ +{ \ + uint64_t *ptr = za + off; \ + TLB(env, useronly_clean_ptr(addr), ptr[BE], ra); \ + TLB(env, useronly_clean_ptr(addr + 8), ptr[!BE], ra); \ +} \ +static inline void VNAME##_v_tlb(CPUARMState *env, void *za, intptr_t off, \ + target_ulong addr, uintptr_t ra) \ +{ \ + HNAME##_tlb(env, za, tile_vslice_offset(off), addr, ra); \ +} + +DO_LD(ld1b, uint8_t, ldub_p, cpu_ldub_data_ra) +DO_LD(ld1h_be, uint16_t, lduw_be_p, cpu_lduw_be_data_ra) +DO_LD(ld1h_le, uint16_t, lduw_le_p, cpu_lduw_le_data_ra) +DO_LD(ld1s_be, uint32_t, ldl_be_p, cpu_ldl_be_data_ra) +DO_LD(ld1s_le, uint32_t, ldl_le_p, cpu_ldl_le_data_ra) +DO_LD(ld1d_be, uint64_t, ldq_be_p, cpu_ldq_be_data_ra) +DO_LD(ld1d_le, uint64_t, ldq_le_p, cpu_ldq_le_data_ra) + +DO_LDQ(sve_ld1qq_be, sme_ld1q_be, 1, ldq_be_p, cpu_ldq_be_data_ra) +DO_LDQ(sve_ld1qq_le, sme_ld1q_le, 0, ldq_le_p, cpu_ldq_le_data_ra) + +DO_ST(st1b, uint8_t, stb_p, cpu_stb_data_ra) +DO_ST(st1h_be, uint16_t, stw_be_p, cpu_stw_be_data_ra) +DO_ST(st1h_le, uint16_t, stw_le_p, cpu_stw_le_data_ra) +DO_ST(st1s_be, uint32_t, stl_be_p, cpu_stl_be_data_ra) +DO_ST(st1s_le, uint32_t, stl_le_p, cpu_stl_le_data_ra) +DO_ST(st1d_be, uint64_t, stq_be_p, cpu_stq_be_data_ra) +DO_ST(st1d_le, uint64_t, stq_le_p, cpu_stq_le_data_ra) + +DO_STQ(sve_st1qq_be, sme_st1q_be, 1, stq_be_p, cpu_stq_be_data_ra) +DO_STQ(sve_st1qq_le, sme_st1q_le, 0, stq_le_p, cpu_stq_le_data_ra) + +#undef DO_LD +#undef DO_ST +#undef DO_LDQ +#undef DO_STQ + +/* + * Common helper for all contiguous predicated loads. + */ + +static inline QEMU_ALWAYS_INLINE +void sme_ld1(CPUARMState *env, void *za, uint64_t *vg, + const target_ulong addr, uint32_t desc, const uintptr_t ra, + const int esz, uint32_t mtedesc, bool vertical, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn, + ClearFn *clr_fn, + CopyFn *cpy_fn) +{ + const intptr_t reg_max = simd_oprsz(desc); + const intptr_t esize = 1 << esz; + intptr_t reg_off, reg_last; + SVEContLdSt info; + void *host; + int flags; + + /* Find the active elements. */ + if (!sve_cont_ldst_elements(&info, addr, vg, reg_max, esz, esize)) { + /* The entire predicate was false; no load occurs. */ + clr_fn(za, 0, reg_max); + return; + } + + /* Probe the page(s). Exit with exception for any invalid page. */ + sve_cont_ldst_pages(&info, FAULT_ALL, env, addr, MMU_DATA_LOAD, ra); + + /* Handle watchpoints for all active elements. */ + sve_cont_ldst_watchpoints(&info, env, vg, addr, esize, esize, + BP_MEM_READ, ra); + + /* + * Handle mte checks for all active elements. + * Since TBI must be set for MTE, !mtedesc => !mte_active. + */ + if (mtedesc) { + sve_cont_ldst_mte_check(&info, env, vg, addr, esize, esize, + mtedesc, ra); + } + + flags = info.page[0].flags | info.page[1].flags; + if (unlikely(flags != 0)) { +#ifdef CONFIG_USER_ONLY + g_assert_not_reached(); +#else + /* + * At least one page includes MMIO. + * Any bus operation can fail with cpu_transaction_failed, + * which for ARM will raise SyncExternal. Perform the load + * into scratch memory to preserve register state until the end. + */ + ARMVectorReg scratch = { }; + + reg_off = info.reg_off_first[0]; + reg_last = info.reg_off_last[1]; + if (reg_last < 0) { + reg_last = info.reg_off_split; + if (reg_last < 0) { + reg_last = info.reg_off_last[0]; + } + } + + do { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + tlb_fn(env, &scratch, reg_off, addr + reg_off, ra); + } + reg_off += esize; + } while (reg_off & 63); + } while (reg_off <= reg_last); + + cpy_fn(za, &scratch, reg_max); + return; +#endif + } + + /* The entire operation is in RAM, on valid pages. */ + + reg_off = info.reg_off_first[0]; + reg_last = info.reg_off_last[0]; + host = info.page[0].host; + + if (!vertical) { + memset(za, 0, reg_max); + } else if (reg_off) { + clr_fn(za, 0, reg_off); + } + + while (reg_off <= reg_last) { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + host_fn(za, reg_off, host + reg_off); + } else if (vertical) { + clr_fn(za, reg_off, esize); + } + reg_off += esize; + } while (reg_off <= reg_last && (reg_off & 63)); + } + + /* + * Use the slow path to manage the cross-page misalignment. + * But we know this is RAM and cannot trap. + */ + reg_off = info.reg_off_split; + if (unlikely(reg_off >= 0)) { + tlb_fn(env, za, reg_off, addr + reg_off, ra); + } + + reg_off = info.reg_off_first[1]; + if (unlikely(reg_off >= 0)) { + reg_last = info.reg_off_last[1]; + host = info.page[1].host; + + do { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + host_fn(za, reg_off, host + reg_off); + } else if (vertical) { + clr_fn(za, reg_off, esize); + } + reg_off += esize; + } while (reg_off & 63); + } while (reg_off <= reg_last); + } +} + +static inline QEMU_ALWAYS_INLINE +void sme_ld1_mte(CPUARMState *env, void *za, uint64_t *vg, + target_ulong addr, uint32_t desc, uintptr_t ra, + const int esz, bool vertical, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn, + ClearFn *clr_fn, + CopyFn *cpy_fn) +{ + uint32_t mtedesc = desc >> (SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + int bit55 = extract64(addr, 55, 1); + + /* Remove mtedesc from the normal sve descriptor. */ + desc = extract32(desc, 0, SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + + /* Perform gross MTE suppression early. */ + if (!tbi_check(desc, bit55) || + tcma_check(desc, bit55, allocation_tag_from_addr(addr))) { + mtedesc = 0; + } + + sme_ld1(env, za, vg, addr, desc, ra, esz, mtedesc, vertical, + host_fn, tlb_fn, clr_fn, cpy_fn); +} + +#define DO_LD(L, END, ESZ) \ +void HELPER(sme_ld1##L##END##_h)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_ld1(env, za, vg, addr, desc, GETPC(), ESZ, 0, false, \ + sve_ld1##L##L##END##_host, sve_ld1##L##L##END##_tlb, \ + clear_horizontal, copy_horizontal); \ +} \ +void HELPER(sme_ld1##L##END##_v)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_ld1(env, za, vg, addr, desc, GETPC(), ESZ, 0, true, \ + sme_ld1##L##END##_v_host, sme_ld1##L##END##_v_tlb, \ + clear_vertical_##L, copy_vertical_##L); \ +} \ +void HELPER(sme_ld1##L##END##_h_mte)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_ld1_mte(env, za, vg, addr, desc, GETPC(), ESZ, false, \ + sve_ld1##L##L##END##_host, sve_ld1##L##L##END##_tlb, \ + clear_horizontal, copy_horizontal); \ +} \ +void HELPER(sme_ld1##L##END##_v_mte)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_ld1_mte(env, za, vg, addr, desc, GETPC(), ESZ, true, \ + sme_ld1##L##END##_v_host, sme_ld1##L##END##_v_tlb, \ + clear_vertical_##L, copy_vertical_##L); \ +} + +DO_LD(b, , MO_8) +DO_LD(h, _be, MO_16) +DO_LD(h, _le, MO_16) +DO_LD(s, _be, MO_32) +DO_LD(s, _le, MO_32) +DO_LD(d, _be, MO_64) +DO_LD(d, _le, MO_64) +DO_LD(q, _be, MO_128) +DO_LD(q, _le, MO_128) + +#undef DO_LD + +/* + * Common helper for all contiguous predicated stores. + */ + +static inline QEMU_ALWAYS_INLINE +void sme_st1(CPUARMState *env, void *za, uint64_t *vg, + const target_ulong addr, uint32_t desc, const uintptr_t ra, + const int esz, uint32_t mtedesc, bool vertical, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn) +{ + const intptr_t reg_max = simd_oprsz(desc); + const intptr_t esize = 1 << esz; + intptr_t reg_off, reg_last; + SVEContLdSt info; + void *host; + int flags; + + /* Find the active elements. */ + if (!sve_cont_ldst_elements(&info, addr, vg, reg_max, esz, esize)) { + /* The entire predicate was false; no store occurs. */ + return; + } + + /* Probe the page(s). Exit with exception for any invalid page. */ + sve_cont_ldst_pages(&info, FAULT_ALL, env, addr, MMU_DATA_STORE, ra); + + /* Handle watchpoints for all active elements. */ + sve_cont_ldst_watchpoints(&info, env, vg, addr, esize, esize, + BP_MEM_WRITE, ra); + + /* + * Handle mte checks for all active elements. + * Since TBI must be set for MTE, !mtedesc => !mte_active. + */ + if (mtedesc) { + sve_cont_ldst_mte_check(&info, env, vg, addr, esize, esize, + mtedesc, ra); + } + + flags = info.page[0].flags | info.page[1].flags; + if (unlikely(flags != 0)) { +#ifdef CONFIG_USER_ONLY + g_assert_not_reached(); +#else + /* + * At least one page includes MMIO. + * Any bus operation can fail with cpu_transaction_failed, + * which for ARM will raise SyncExternal. We cannot avoid + * this fault and will leave with the store incomplete. + */ + reg_off = info.reg_off_first[0]; + reg_last = info.reg_off_last[1]; + if (reg_last < 0) { + reg_last = info.reg_off_split; + if (reg_last < 0) { + reg_last = info.reg_off_last[0]; + } + } + + do { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + tlb_fn(env, za, reg_off, addr + reg_off, ra); + } + reg_off += esize; + } while (reg_off & 63); + } while (reg_off <= reg_last); + return; +#endif + } + + reg_off = info.reg_off_first[0]; + reg_last = info.reg_off_last[0]; + host = info.page[0].host; + + while (reg_off <= reg_last) { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + host_fn(za, reg_off, host + reg_off); + } + reg_off += 1 << esz; + } while (reg_off <= reg_last && (reg_off & 63)); + } + + /* + * Use the slow path to manage the cross-page misalignment. + * But we know this is RAM and cannot trap. + */ + reg_off = info.reg_off_split; + if (unlikely(reg_off >= 0)) { + tlb_fn(env, za, reg_off, addr + reg_off, ra); + } + + reg_off = info.reg_off_first[1]; + if (unlikely(reg_off >= 0)) { + reg_last = info.reg_off_last[1]; + host = info.page[1].host; + + do { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + host_fn(za, reg_off, host + reg_off); + } + reg_off += 1 << esz; + } while (reg_off & 63); + } while (reg_off <= reg_last); + } +} + +static inline QEMU_ALWAYS_INLINE +void sme_st1_mte(CPUARMState *env, void *za, uint64_t *vg, target_ulong addr, + uint32_t desc, uintptr_t ra, int esz, bool vertical, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn) +{ + uint32_t mtedesc = desc >> (SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + int bit55 = extract64(addr, 55, 1); + + /* Remove mtedesc from the normal sve descriptor. */ + desc = extract32(desc, 0, SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + + /* Perform gross MTE suppression early. */ + if (!tbi_check(desc, bit55) || + tcma_check(desc, bit55, allocation_tag_from_addr(addr))) { + mtedesc = 0; + } + + sme_st1(env, za, vg, addr, desc, ra, esz, mtedesc, + vertical, host_fn, tlb_fn); +} + +#define DO_ST(L, END, ESZ) \ +void HELPER(sme_st1##L##END##_h)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_st1(env, za, vg, addr, desc, GETPC(), ESZ, 0, false, \ + sve_st1##L##L##END##_host, sve_st1##L##L##END##_tlb); \ +} \ +void HELPER(sme_st1##L##END##_v)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_st1(env, za, vg, addr, desc, GETPC(), ESZ, 0, true, \ + sme_st1##L##END##_v_host, sme_st1##L##END##_v_tlb); \ +} \ +void HELPER(sme_st1##L##END##_h_mte)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_st1_mte(env, za, vg, addr, desc, GETPC(), ESZ, false, \ + sve_st1##L##L##END##_host, sve_st1##L##L##END##_tlb); \ +} \ +void HELPER(sme_st1##L##END##_v_mte)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_st1_mte(env, za, vg, addr, desc, GETPC(), ESZ, true, \ + sme_st1##L##END##_v_host, sme_st1##L##END##_v_tlb); \ +} + +DO_ST(b, , MO_8) +DO_ST(h, _be, MO_16) +DO_ST(h, _le, MO_16) +DO_ST(s, _be, MO_32) +DO_ST(s, _le, MO_32) +DO_ST(d, _be, MO_64) +DO_ST(d, _le, MO_64) +DO_ST(q, _be, MO_128) +DO_ST(q, _le, MO_128) + +#undef DO_ST diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ static bool trans_MOVA(DisasContext *s, arg_MOVA *a) return true; } + +static bool trans_LDST1(DisasContext *s, arg_LDST1 *a) +{ + typedef void GenLdSt1(TCGv_env, TCGv_ptr, TCGv_ptr, TCGv, TCGv_i32); + + /* + * Indexed by [esz][be][v][mte][st], which is (except for load/store) + * also the order in which the elements appear in the function names, + * and so how we must concatenate the pieces. + */ + +#define FN_LS(F) { gen_helper_sme_ld1##F, gen_helper_sme_st1##F } +#define FN_MTE(F) { FN_LS(F), FN_LS(F##_mte) } +#define FN_HV(F) { FN_MTE(F##_h), FN_MTE(F##_v) } +#define FN_END(L, B) { FN_HV(L), FN_HV(B) } + + static GenLdSt1 * const fns[5][2][2][2][2] = { + FN_END(b, b), + FN_END(h_le, h_be), + FN_END(s_le, s_be), + FN_END(d_le, d_be), + FN_END(q_le, q_be), + }; + +#undef FN_LS +#undef FN_MTE +#undef FN_HV +#undef FN_END + + TCGv_ptr t_za, t_pg; + TCGv_i64 addr; + int svl, desc = 0; + bool be = s->be_data == MO_BE; + bool mte = s->mte_active[0]; + + if (!dc_isar_feature(aa64_sme, s)) { + return false; + } + if (!sme_smza_enabled_check(s)) { + return true; + } + + t_za = get_tile_rowcol(s, a->esz, a->rs, a->za_imm, a->v); + t_pg = pred_full_reg_ptr(s, a->pg); + addr = tcg_temp_new_i64(); + + tcg_gen_shli_i64(addr, cpu_reg(s, a->rm), a->esz); + tcg_gen_add_i64(addr, addr, cpu_reg_sp(s, a->rn)); + + if (mte) { + desc = FIELD_DP32(desc, MTEDESC, MIDX, get_mem_index(s)); + desc = FIELD_DP32(desc, MTEDESC, TBI, s->tbid); + desc = FIELD_DP32(desc, MTEDESC, TCMA, s->tcma); + desc = FIELD_DP32(desc, MTEDESC, WRITE, a->st); + desc = FIELD_DP32(desc, MTEDESC, SIZEM1, (1 << a->esz) - 1); + desc <<= SVE_MTEDESC_SHIFT; + } else { + addr = clean_data_tbi(s, addr); + } + svl = streaming_vec_reg_size(s); + desc = simd_desc(svl, svl, desc); + + fns[a->esz][be][a->v][mte][a->st](cpu_env, t_za, t_pg, addr, + tcg_constant_i32(desc)); + + tcg_temp_free_ptr(t_za); + tcg_temp_free_ptr(t_pg); + tcg_temp_free_i64(addr); + return true; +} -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Add a TCGv_ptr base argument, which will be cpu_env for SVE. We will reuse this for SME save and restore array insns. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-22-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.h | 3 +++ target/arm/translate-sve.c | 48 ++++++++++++++++++++++++++++---------- 2 files changed, 39 insertions(+), 12 deletions(-) diff --git a/target/arm/translate-a64.h b/target/arm/translate-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.h +++ b/target/arm/translate-a64.h @@ -XXX,XX +XXX,XX @@ void gen_gvec_xar(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, uint32_t rm_ofs, int64_t shift, uint32_t opr_sz, uint32_t max_sz); +void gen_sve_ldr(DisasContext *s, TCGv_ptr, int vofs, int len, int rn, int imm); +void gen_sve_str(DisasContext *s, TCGv_ptr, int vofs, int len, int rn, int imm); + #endif /* TARGET_ARM_TRANSLATE_A64_H */ 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 @@ TRANS_FEAT(UCVTF_dd, aa64_sve, gen_gvec_fpst_arg_zpz, * The load should begin at the address Rn + IMM. */ -static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) +void gen_sve_ldr(DisasContext *s, TCGv_ptr base, int vofs, + int len, int rn, int imm) { int len_align = QEMU_ALIGN_DOWN(len, 8); int len_remain = len % 8; @@ -XXX,XX +XXX,XX @@ static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) t0 = tcg_temp_new_i64(); for (i = 0; i < len_align; i += 8) { tcg_gen_qemu_ld_i64(t0, clean_addr, midx, MO_LEUQ); - tcg_gen_st_i64(t0, cpu_env, vofs + i); + tcg_gen_st_i64(t0, base, vofs + i); tcg_gen_addi_i64(clean_addr, clean_addr, 8); } tcg_temp_free_i64(t0); @@ -XXX,XX +XXX,XX @@ static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) clean_addr = new_tmp_a64_local(s); tcg_gen_mov_i64(clean_addr, t0); + if (base != cpu_env) { + TCGv_ptr b = tcg_temp_local_new_ptr(); + tcg_gen_mov_ptr(b, base); + base = b; + } + gen_set_label(loop); t0 = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) tcg_gen_addi_i64(clean_addr, clean_addr, 8); tp = tcg_temp_new_ptr(); - tcg_gen_add_ptr(tp, cpu_env, i); + tcg_gen_add_ptr(tp, base, i); tcg_gen_addi_ptr(i, i, 8); tcg_gen_st_i64(t0, tp, vofs); tcg_temp_free_ptr(tp); @@ -XXX,XX +XXX,XX @@ static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) tcg_gen_brcondi_ptr(TCG_COND_LTU, i, len_align, loop); tcg_temp_free_ptr(i); + + if (base != cpu_env) { + tcg_temp_free_ptr(base); + assert(len_remain == 0); + } } /* @@ -XXX,XX +XXX,XX @@ static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) default: g_assert_not_reached(); } - tcg_gen_st_i64(t0, cpu_env, vofs + len_align); + tcg_gen_st_i64(t0, base, vofs + len_align); tcg_temp_free_i64(t0); } } /* Similarly for stores. */ -static void do_str(DisasContext *s, uint32_t vofs, int len, int rn, int imm) +void gen_sve_str(DisasContext *s, TCGv_ptr base, int vofs, + int len, int rn, int imm) { int len_align = QEMU_ALIGN_DOWN(len, 8); int len_remain = len % 8; @@ -XXX,XX +XXX,XX @@ static void do_str(DisasContext *s, uint32_t vofs, int len, int rn, int imm) t0 = tcg_temp_new_i64(); for (i = 0; i < len_align; i += 8) { - tcg_gen_ld_i64(t0, cpu_env, vofs + i); + tcg_gen_ld_i64(t0, base, vofs + i); tcg_gen_qemu_st_i64(t0, clean_addr, midx, MO_LEUQ); tcg_gen_addi_i64(clean_addr, clean_addr, 8); } @@ -XXX,XX +XXX,XX @@ static void do_str(DisasContext *s, uint32_t vofs, int len, int rn, int imm) clean_addr = new_tmp_a64_local(s); tcg_gen_mov_i64(clean_addr, t0); + if (base != cpu_env) { + TCGv_ptr b = tcg_temp_local_new_ptr(); + tcg_gen_mov_ptr(b, base); + base = b; + } + gen_set_label(loop); t0 = tcg_temp_new_i64(); tp = tcg_temp_new_ptr(); - tcg_gen_add_ptr(tp, cpu_env, i); + tcg_gen_add_ptr(tp, base, i); tcg_gen_ld_i64(t0, tp, vofs); tcg_gen_addi_ptr(i, i, 8); tcg_temp_free_ptr(tp); @@ -XXX,XX +XXX,XX @@ static void do_str(DisasContext *s, uint32_t vofs, int len, int rn, int imm) tcg_gen_brcondi_ptr(TCG_COND_LTU, i, len_align, loop); tcg_temp_free_ptr(i); + + if (base != cpu_env) { + tcg_temp_free_ptr(base); + assert(len_remain == 0); + } } /* Predicate register stores can be any multiple of 2. */ if (len_remain) { t0 = tcg_temp_new_i64(); - tcg_gen_ld_i64(t0, cpu_env, vofs + len_align); + tcg_gen_ld_i64(t0, base, vofs + len_align); switch (len_remain) { case 2: @@ -XXX,XX +XXX,XX @@ static bool trans_LDR_zri(DisasContext *s, arg_rri *a) if (sve_access_check(s)) { int size = vec_full_reg_size(s); int off = vec_full_reg_offset(s, a->rd); - do_ldr(s, off, size, a->rn, a->imm * size); + gen_sve_ldr(s, cpu_env, off, size, a->rn, a->imm * size); } return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_LDR_pri(DisasContext *s, arg_rri *a) if (sve_access_check(s)) { int size = pred_full_reg_size(s); int off = pred_full_reg_offset(s, a->rd); - do_ldr(s, off, size, a->rn, a->imm * size); + gen_sve_ldr(s, cpu_env, off, size, a->rn, a->imm * size); } return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_STR_zri(DisasContext *s, arg_rri *a) if (sve_access_check(s)) { int size = vec_full_reg_size(s); int off = vec_full_reg_offset(s, a->rd); - do_str(s, off, size, a->rn, a->imm * size); + gen_sve_str(s, cpu_env, off, size, a->rn, a->imm * size); } return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_STR_pri(DisasContext *s, arg_rri *a) if (sve_access_check(s)) { int size = pred_full_reg_size(s); int off = pred_full_reg_offset(s, a->rd); - do_str(s, off, size, a->rn, a->imm * size); + gen_sve_str(s, cpu_env, off, size, a->rn, a->imm * size); } return true; } -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> We can reuse the SVE functions for LDR and STR, passing in the base of the ZA vector and a zero offset. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-23-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme.decode | 7 +++++++ target/arm/translate-sme.c | 24 ++++++++++++++++++++++++ 2 files changed, 31 insertions(+) diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ LDST1 1110000 0 esz:2 st:1 rm:5 v:1 .. pg:3 rn:5 0 za_imm:4 \ &ldst rs=%mova_rs LDST1 1110000 111 st:1 rm:5 v:1 .. pg:3 rn:5 0 za_imm:4 \ &ldst esz=4 rs=%mova_rs + +&ldstr rv rn imm +@ldstr ....... ... . ...... .. ... rn:5 . imm:4 \ + &ldstr rv=%mova_rs + +LDR 1110000 100 0 000000 .. 000 ..... 0 .... @ldstr +STR 1110000 100 1 000000 .. 000 ..... 0 .... @ldstr diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ static bool trans_LDST1(DisasContext *s, arg_LDST1 *a) tcg_temp_free_i64(addr); return true; } + +typedef void GenLdStR(DisasContext *, TCGv_ptr, int, int, int, int); + +static bool do_ldst_r(DisasContext *s, arg_ldstr *a, GenLdStR *fn) +{ + int svl = streaming_vec_reg_size(s); + int imm = a->imm; + TCGv_ptr base; + + if (!sme_za_enabled_check(s)) { + return true; + } + + /* ZA[n] equates to ZA0H.B[n]. */ + base = get_tile_rowcol(s, MO_8, a->rv, imm, false); + + fn(s, base, 0, svl, a->rn, imm * svl); + + tcg_temp_free_ptr(base); + return true; +} + +TRANS_FEAT(LDR, aa64_sme, do_ldst_r, a, gen_sve_ldr) +TRANS_FEAT(STR, aa64_sme, do_ldst_r, a, gen_sve_str) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-24-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 5 +++ target/arm/sme.decode | 11 +++++ target/arm/sme_helper.c | 90 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-sme.c | 31 +++++++++++++ 4 files changed, 137 insertions(+) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_5(sme_st1q_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i DEF_HELPER_FLAGS_5(sme_st1q_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_5(sme_st1q_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_5(sme_st1q_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_addha_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(sme_addva_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(sme_addha_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(sme_addva_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ LDST1 1110000 111 st:1 rm:5 v:1 .. pg:3 rn:5 0 za_imm:4 \ LDR 1110000 100 0 000000 .. 000 ..... 0 .... @ldstr STR 1110000 100 1 000000 .. 000 ..... 0 .... @ldstr + +### SME Add Vector to Array + +&adda zad zn pm pn +@adda_32 ........ .. ..... . pm:3 pn:3 zn:5 ... zad:2 &adda +@adda_64 ........ .. ..... . pm:3 pn:3 zn:5 .. zad:3 &adda + +ADDHA_s 11000000 10 01000 0 ... ... ..... 000 .. @adda_32 +ADDVA_s 11000000 10 01000 1 ... ... ..... 000 .. @adda_32 +ADDHA_d 11000000 11 01000 0 ... ... ..... 00 ... @adda_64 +ADDVA_d 11000000 11 01000 1 ... ... ..... 00 ... @adda_64 diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ DO_ST(q, _be, MO_128) DO_ST(q, _le, MO_128) #undef DO_ST + +void HELPER(sme_addha_s)(void *vzda, void *vzn, void *vpn, + void *vpm, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_oprsz(desc) / 4; + uint64_t *pn = vpn, *pm = vpm; + uint32_t *zda = vzda, *zn = vzn; + + for (row = 0; row < oprsz; ) { + uint64_t pa = pn[row >> 4]; + do { + if (pa & 1) { + for (col = 0; col < oprsz; ) { + uint64_t pb = pm[col >> 4]; + do { + if (pb & 1) { + zda[tile_vslice_index(row) + H4(col)] += zn[H4(col)]; + } + pb >>= 4; + } while (++col & 15); + } + } + pa >>= 4; + } while (++row & 15); + } +} + +void HELPER(sme_addha_d)(void *vzda, void *vzn, void *vpn, + void *vpm, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_oprsz(desc) / 8; + uint8_t *pn = vpn, *pm = vpm; + uint64_t *zda = vzda, *zn = vzn; + + for (row = 0; row < oprsz; ++row) { + if (pn[H1(row)] & 1) { + for (col = 0; col < oprsz; ++col) { + if (pm[H1(col)] & 1) { + zda[tile_vslice_index(row) + col] += zn[col]; + } + } + } + } +} + +void HELPER(sme_addva_s)(void *vzda, void *vzn, void *vpn, + void *vpm, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_oprsz(desc) / 4; + uint64_t *pn = vpn, *pm = vpm; + uint32_t *zda = vzda, *zn = vzn; + + for (row = 0; row < oprsz; ) { + uint64_t pa = pn[row >> 4]; + do { + if (pa & 1) { + uint32_t zn_row = zn[H4(row)]; + for (col = 0; col < oprsz; ) { + uint64_t pb = pm[col >> 4]; + do { + if (pb & 1) { + zda[tile_vslice_index(row) + H4(col)] += zn_row; + } + pb >>= 4; + } while (++col & 15); + } + } + pa >>= 4; + } while (++row & 15); + } +} + +void HELPER(sme_addva_d)(void *vzda, void *vzn, void *vpn, + void *vpm, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_oprsz(desc) / 8; + uint8_t *pn = vpn, *pm = vpm; + uint64_t *zda = vzda, *zn = vzn; + + for (row = 0; row < oprsz; ++row) { + if (pn[H1(row)] & 1) { + uint64_t zn_row = zn[row]; + for (col = 0; col < oprsz; ++col) { + if (pm[H1(col)] & 1) { + zda[tile_vslice_index(row) + col] += zn_row; + } + } + } + } +} diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ static bool do_ldst_r(DisasContext *s, arg_ldstr *a, GenLdStR *fn) TRANS_FEAT(LDR, aa64_sme, do_ldst_r, a, gen_sve_ldr) TRANS_FEAT(STR, aa64_sme, do_ldst_r, a, gen_sve_str) + +static bool do_adda(DisasContext *s, arg_adda *a, MemOp esz, + gen_helper_gvec_4 *fn) +{ + int svl = streaming_vec_reg_size(s); + uint32_t desc = simd_desc(svl, svl, 0); + TCGv_ptr za, zn, pn, pm; + + if (!sme_smza_enabled_check(s)) { + return true; + } + + /* Sum XZR+zad to find ZAd. */ + za = get_tile_rowcol(s, esz, 31, a->zad, false); + zn = vec_full_reg_ptr(s, a->zn); + pn = pred_full_reg_ptr(s, a->pn); + pm = pred_full_reg_ptr(s, a->pm); + + fn(za, zn, pn, pm, tcg_constant_i32(desc)); + + tcg_temp_free_ptr(za); + tcg_temp_free_ptr(zn); + tcg_temp_free_ptr(pn); + tcg_temp_free_ptr(pm); + return true; +} + +TRANS_FEAT(ADDHA_s, aa64_sme, do_adda, a, MO_32, gen_helper_sme_addha_s) +TRANS_FEAT(ADDVA_s, aa64_sme, do_adda, a, MO_32, gen_helper_sme_addva_s) +TRANS_FEAT(ADDHA_d, aa64_sme_i16i64, do_adda, a, MO_64, gen_helper_sme_addha_d) +TRANS_FEAT(ADDVA_d, aa64_sme_i16i64, do_adda, a, MO_64, gen_helper_sme_addva_d) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-25-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 5 +++ target/arm/sme.decode | 9 +++++ target/arm/sme_helper.c | 69 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-sme.c | 32 ++++++++++++++++++ 4 files changed, 115 insertions(+) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_5(sme_addha_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sme_addva_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sme_addha_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sme_addva_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_7(sme_fmopa_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_7(sme_fmopa_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, ptr, i32) diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ ADDHA_s 11000000 10 01000 0 ... ... ..... 000 .. @adda_32 ADDVA_s 11000000 10 01000 1 ... ... ..... 000 .. @adda_32 ADDHA_d 11000000 11 01000 0 ... ... ..... 00 ... @adda_64 ADDVA_d 11000000 11 01000 1 ... ... ..... 00 ... @adda_64 + +### SME Outer Product + +&op zad zn zm pm pn sub:bool +@op_32 ........ ... zm:5 pm:3 pn:3 zn:5 sub:1 .. zad:2 &op +@op_64 ........ ... zm:5 pm:3 pn:3 zn:5 sub:1 . zad:3 &op + +FMOPA_s 10000000 100 ..... ... ... ..... . 00 .. @op_32 +FMOPA_d 10000000 110 ..... ... ... ..... . 0 ... @op_64 diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ #include "exec/cpu_ldst.h" #include "exec/exec-all.h" #include "qemu/int128.h" +#include "fpu/softfloat.h" #include "vec_internal.h" #include "sve_ldst_internal.h" @@ -XXX,XX +XXX,XX @@ void HELPER(sme_addva_d)(void *vzda, void *vzn, void *vpn, } } } + +void HELPER(sme_fmopa_s)(void *vza, void *vzn, void *vzm, void *vpn, + void *vpm, void *vst, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_maxsz(desc); + uint32_t neg = simd_data(desc) << 31; + uint16_t *pn = vpn, *pm = vpm; + float_status fpst; + + /* + * Make a copy of float_status because this operation does not + * update the cumulative fp exception status. It also produces + * default nans. + */ + fpst = *(float_status *)vst; + set_default_nan_mode(true, &fpst); + + for (row = 0; row < oprsz; ) { + uint16_t pa = pn[H2(row >> 4)]; + do { + if (pa & 1) { + void *vza_row = vza + tile_vslice_offset(row); + uint32_t n = *(uint32_t *)(vzn + H1_4(row)) ^ neg; + + for (col = 0; col < oprsz; ) { + uint16_t pb = pm[H2(col >> 4)]; + do { + if (pb & 1) { + uint32_t *a = vza_row + H1_4(col); + uint32_t *m = vzm + H1_4(col); + *a = float32_muladd(n, *m, *a, 0, vst); + } + col += 4; + pb >>= 4; + } while (col & 15); + } + } + row += 4; + pa >>= 4; + } while (row & 15); + } +} + +void HELPER(sme_fmopa_d)(void *vza, void *vzn, void *vzm, void *vpn, + void *vpm, void *vst, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_oprsz(desc) / 8; + uint64_t neg = (uint64_t)simd_data(desc) << 63; + uint64_t *za = vza, *zn = vzn, *zm = vzm; + uint8_t *pn = vpn, *pm = vpm; + float_status fpst = *(float_status *)vst; + + set_default_nan_mode(true, &fpst); + + for (row = 0; row < oprsz; ++row) { + if (pn[H1(row)] & 1) { + uint64_t *za_row = &za[tile_vslice_index(row)]; + uint64_t n = zn[row] ^ neg; + + for (col = 0; col < oprsz; ++col) { + if (pm[H1(col)] & 1) { + uint64_t *a = &za_row[col]; + *a = float64_muladd(n, zm[col], *a, 0, &fpst); + } + } + } + } +} diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ TRANS_FEAT(ADDHA_s, aa64_sme, do_adda, a, MO_32, gen_helper_sme_addha_s) TRANS_FEAT(ADDVA_s, aa64_sme, do_adda, a, MO_32, gen_helper_sme_addva_s) TRANS_FEAT(ADDHA_d, aa64_sme_i16i64, do_adda, a, MO_64, gen_helper_sme_addha_d) TRANS_FEAT(ADDVA_d, aa64_sme_i16i64, do_adda, a, MO_64, gen_helper_sme_addva_d) + +static bool do_outprod_fpst(DisasContext *s, arg_op *a, MemOp esz, + gen_helper_gvec_5_ptr *fn) +{ + int svl = streaming_vec_reg_size(s); + uint32_t desc = simd_desc(svl, svl, a->sub); + TCGv_ptr za, zn, zm, pn, pm, fpst; + + if (!sme_smza_enabled_check(s)) { + return true; + } + + /* Sum XZR+zad to find ZAd. */ + za = get_tile_rowcol(s, esz, 31, a->zad, false); + zn = vec_full_reg_ptr(s, a->zn); + zm = vec_full_reg_ptr(s, a->zm); + pn = pred_full_reg_ptr(s, a->pn); + pm = pred_full_reg_ptr(s, a->pm); + fpst = fpstatus_ptr(FPST_FPCR); + + fn(za, zn, zm, pn, pm, fpst, tcg_constant_i32(desc)); + + tcg_temp_free_ptr(za); + tcg_temp_free_ptr(zn); + tcg_temp_free_ptr(pn); + tcg_temp_free_ptr(pm); + tcg_temp_free_ptr(fpst); + return true; +} + +TRANS_FEAT(FMOPA_s, aa64_sme, do_outprod_fpst, a, MO_32, gen_helper_sme_fmopa_s) +TRANS_FEAT(FMOPA_d, aa64_sme_f64f64, do_outprod_fpst, a, MO_64, gen_helper_sme_fmopa_d) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-26-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 2 ++ target/arm/sme.decode | 2 ++ target/arm/sme_helper.c | 56 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-sme.c | 30 ++++++++++++++++++++ 4 files changed, 90 insertions(+) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_7(sme_fmopa_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_7(sme_fmopa_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_bfmopa, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ ADDVA_d 11000000 11 01000 1 ... ... ..... 00 ... @adda_64 FMOPA_s 10000000 100 ..... ... ... ..... . 00 .. @op_32 FMOPA_d 10000000 110 ..... ... ... ..... . 0 ... @op_64 + +BFMOPA 10000001 100 ..... ... ... ..... . 00 .. @op_32 diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(sme_fmopa_d)(void *vza, void *vzn, void *vzm, void *vpn, } } } + +/* + * Alter PAIR as needed for controlling predicates being false, + * and for NEG on an enabled row element. + */ +static inline uint32_t f16mop_adj_pair(uint32_t pair, uint32_t pg, uint32_t neg) +{ + /* + * The pseudocode uses a conditional negate after the conditional zero. + * It is simpler here to unconditionally negate before conditional zero. + */ + pair ^= neg; + if (!(pg & 1)) { + pair &= 0xffff0000u; + } + if (!(pg & 4)) { + pair &= 0x0000ffffu; + } + return pair; +} + +void HELPER(sme_bfmopa)(void *vza, void *vzn, void *vzm, void *vpn, + void *vpm, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_maxsz(desc); + uint32_t neg = simd_data(desc) * 0x80008000u; + uint16_t *pn = vpn, *pm = vpm; + + for (row = 0; row < oprsz; ) { + uint16_t prow = pn[H2(row >> 4)]; + do { + void *vza_row = vza + tile_vslice_offset(row); + uint32_t n = *(uint32_t *)(vzn + H1_4(row)); + + n = f16mop_adj_pair(n, prow, neg); + + for (col = 0; col < oprsz; ) { + uint16_t pcol = pm[H2(col >> 4)]; + do { + if (prow & pcol & 0b0101) { + uint32_t *a = vza_row + H1_4(col); + uint32_t m = *(uint32_t *)(vzm + H1_4(col)); + + m = f16mop_adj_pair(m, pcol, 0); + *a = bfdotadd(*a, n, m); + + col += 4; + pcol >>= 4; + } + } while (col & 15); + } + row += 4; + prow >>= 4; + } while (row & 15); + } +} diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ TRANS_FEAT(ADDVA_s, aa64_sme, do_adda, a, MO_32, gen_helper_sme_addva_s) TRANS_FEAT(ADDHA_d, aa64_sme_i16i64, do_adda, a, MO_64, gen_helper_sme_addha_d) TRANS_FEAT(ADDVA_d, aa64_sme_i16i64, do_adda, a, MO_64, gen_helper_sme_addva_d) +static bool do_outprod(DisasContext *s, arg_op *a, MemOp esz, + gen_helper_gvec_5 *fn) +{ + int svl = streaming_vec_reg_size(s); + uint32_t desc = simd_desc(svl, svl, a->sub); + TCGv_ptr za, zn, zm, pn, pm; + + if (!sme_smza_enabled_check(s)) { + return true; + } + + /* Sum XZR+zad to find ZAd. */ + za = get_tile_rowcol(s, esz, 31, a->zad, false); + zn = vec_full_reg_ptr(s, a->zn); + zm = vec_full_reg_ptr(s, a->zm); + pn = pred_full_reg_ptr(s, a->pn); + pm = pred_full_reg_ptr(s, a->pm); + + fn(za, zn, zm, pn, pm, tcg_constant_i32(desc)); + + tcg_temp_free_ptr(za); + tcg_temp_free_ptr(zn); + tcg_temp_free_ptr(pn); + tcg_temp_free_ptr(pm); + return true; +} + static bool do_outprod_fpst(DisasContext *s, arg_op *a, MemOp esz, gen_helper_gvec_5_ptr *fn) { @@ -XXX,XX +XXX,XX @@ static bool do_outprod_fpst(DisasContext *s, arg_op *a, MemOp esz, TRANS_FEAT(FMOPA_s, aa64_sme, do_outprod_fpst, a, MO_32, gen_helper_sme_fmopa_s) TRANS_FEAT(FMOPA_d, aa64_sme_f64f64, do_outprod_fpst, a, MO_64, gen_helper_sme_fmopa_d) + +/* TODO: FEAT_EBF16 */ +TRANS_FEAT(BFMOPA, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_bfmopa) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-27-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 2 ++ target/arm/sme.decode | 1 + target/arm/sme_helper.c | 74 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-sme.c | 1 + 4 files changed, 78 insertions(+) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_5(sme_addva_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sme_addha_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sme_addva_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_7(sme_fmopa_h, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_7(sme_fmopa_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_7(sme_fmopa_d, TCG_CALL_NO_RWG, diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ FMOPA_s 10000000 100 ..... ... ... ..... . 00 .. @op_32 FMOPA_d 10000000 110 ..... ... ... ..... . 0 ... @op_64 BFMOPA 10000001 100 ..... ... ... ..... . 00 .. @op_32 +FMOPA_h 10000001 101 ..... ... ... ..... . 00 .. @op_32 diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ static inline uint32_t f16mop_adj_pair(uint32_t pair, uint32_t pg, uint32_t neg) return pair; } +static float32 f16_dotadd(float32 sum, uint32_t e1, uint32_t e2, + float_status *s_std, float_status *s_odd) +{ + float64 e1r = float16_to_float64(e1 & 0xffff, true, s_std); + float64 e1c = float16_to_float64(e1 >> 16, true, s_std); + float64 e2r = float16_to_float64(e2 & 0xffff, true, s_std); + float64 e2c = float16_to_float64(e2 >> 16, true, s_std); + float64 t64; + float32 t32; + + /* + * The ARM pseudocode function FPDot performs both multiplies + * and the add with a single rounding operation. Emulate this + * by performing the first multiply in round-to-odd, then doing + * the second multiply as fused multiply-add, and rounding to + * float32 all in one step. + */ + t64 = float64_mul(e1r, e2r, s_odd); + t64 = float64r32_muladd(e1c, e2c, t64, 0, s_std); + + /* This conversion is exact, because we've already rounded. */ + t32 = float64_to_float32(t64, s_std); + + /* The final accumulation step is not fused. */ + return float32_add(sum, t32, s_std); +} + +void HELPER(sme_fmopa_h)(void *vza, void *vzn, void *vzm, void *vpn, + void *vpm, void *vst, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_maxsz(desc); + uint32_t neg = simd_data(desc) * 0x80008000u; + uint16_t *pn = vpn, *pm = vpm; + float_status fpst_odd, fpst_std; + + /* + * Make a copy of float_status because this operation does not + * update the cumulative fp exception status. It also produces + * default nans. Make a second copy with round-to-odd -- see above. + */ + fpst_std = *(float_status *)vst; + set_default_nan_mode(true, &fpst_std); + fpst_odd = fpst_std; + set_float_rounding_mode(float_round_to_odd, &fpst_odd); + + for (row = 0; row < oprsz; ) { + uint16_t prow = pn[H2(row >> 4)]; + do { + void *vza_row = vza + tile_vslice_offset(row); + uint32_t n = *(uint32_t *)(vzn + H1_4(row)); + + n = f16mop_adj_pair(n, prow, neg); + + for (col = 0; col < oprsz; ) { + uint16_t pcol = pm[H2(col >> 4)]; + do { + if (prow & pcol & 0b0101) { + uint32_t *a = vza_row + H1_4(col); + uint32_t m = *(uint32_t *)(vzm + H1_4(col)); + + m = f16mop_adj_pair(m, pcol, 0); + *a = f16_dotadd(*a, n, m, &fpst_std, &fpst_odd); + + col += 4; + pcol >>= 4; + } + } while (col & 15); + } + row += 4; + prow >>= 4; + } while (row & 15); + } +} + void HELPER(sme_bfmopa)(void *vza, void *vzn, void *vzm, void *vpn, void *vpm, uint32_t desc) { diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ static bool do_outprod_fpst(DisasContext *s, arg_op *a, MemOp esz, return true; } +TRANS_FEAT(FMOPA_h, aa64_sme, do_outprod_fpst, a, MO_32, gen_helper_sme_fmopa_h) TRANS_FEAT(FMOPA_s, aa64_sme, do_outprod_fpst, a, MO_32, gen_helper_sme_fmopa_s) TRANS_FEAT(FMOPA_d, aa64_sme_f64f64, do_outprod_fpst, a, MO_64, gen_helper_sme_fmopa_d) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This is SMOPA, SUMOPA, USMOPA_s, UMOPA, for both Int8 and Int16. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-28-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 16 ++++++++ target/arm/sme.decode | 10 +++++ target/arm/sme_helper.c | 82 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-sme.c | 10 +++++ 4 files changed, 118 insertions(+) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_7(sme_fmopa_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_6(sme_bfmopa, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_smopa_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_umopa_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_sumopa_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_usmopa_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_smopa_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_umopa_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_sumopa_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_usmopa_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ FMOPA_d 10000000 110 ..... ... ... ..... . 0 ... @op_64 BFMOPA 10000001 100 ..... ... ... ..... . 00 .. @op_32 FMOPA_h 10000001 101 ..... ... ... ..... . 00 .. @op_32 + +SMOPA_s 1010000 0 10 0 ..... ... ... ..... . 00 .. @op_32 +SUMOPA_s 1010000 0 10 1 ..... ... ... ..... . 00 .. @op_32 +USMOPA_s 1010000 1 10 0 ..... ... ... ..... . 00 .. @op_32 +UMOPA_s 1010000 1 10 1 ..... ... ... ..... . 00 .. @op_32 + +SMOPA_d 1010000 0 11 0 ..... ... ... ..... . 0 ... @op_64 +SUMOPA_d 1010000 0 11 1 ..... ... ... ..... . 0 ... @op_64 +USMOPA_d 1010000 1 11 0 ..... ... ... ..... . 0 ... @op_64 +UMOPA_d 1010000 1 11 1 ..... ... ... ..... . 0 ... @op_64 diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(sme_bfmopa)(void *vza, void *vzn, void *vzm, void *vpn, } while (row & 15); } } + +typedef uint64_t IMOPFn(uint64_t, uint64_t, uint64_t, uint8_t, bool); + +static inline void do_imopa(uint64_t *za, uint64_t *zn, uint64_t *zm, + uint8_t *pn, uint8_t *pm, + uint32_t desc, IMOPFn *fn) +{ + intptr_t row, col, oprsz = simd_oprsz(desc) / 8; + bool neg = simd_data(desc); + + for (row = 0; row < oprsz; ++row) { + uint8_t pa = pn[H1(row)]; + uint64_t *za_row = &za[tile_vslice_index(row)]; + uint64_t n = zn[row]; + + for (col = 0; col < oprsz; ++col) { + uint8_t pb = pm[H1(col)]; + uint64_t *a = &za_row[col]; + + *a = fn(n, zm[col], *a, pa & pb, neg); + } + } +} + +#define DEF_IMOP_32(NAME, NTYPE, MTYPE) \ +static uint64_t NAME(uint64_t n, uint64_t m, uint64_t a, uint8_t p, bool neg) \ +{ \ + uint32_t sum0 = 0, sum1 = 0; \ + /* Apply P to N as a mask, making the inactive elements 0. */ \ + n &= expand_pred_b(p); \ + sum0 += (NTYPE)(n >> 0) * (MTYPE)(m >> 0); \ + sum0 += (NTYPE)(n >> 8) * (MTYPE)(m >> 8); \ + sum0 += (NTYPE)(n >> 16) * (MTYPE)(m >> 16); \ + sum0 += (NTYPE)(n >> 24) * (MTYPE)(m >> 24); \ + sum1 += (NTYPE)(n >> 32) * (MTYPE)(m >> 32); \ + sum1 += (NTYPE)(n >> 40) * (MTYPE)(m >> 40); \ + sum1 += (NTYPE)(n >> 48) * (MTYPE)(m >> 48); \ + sum1 += (NTYPE)(n >> 56) * (MTYPE)(m >> 56); \ + if (neg) { \ + sum0 = (uint32_t)a - sum0, sum1 = (uint32_t)(a >> 32) - sum1; \ + } else { \ + sum0 = (uint32_t)a + sum0, sum1 = (uint32_t)(a >> 32) + sum1; \ + } \ + return ((uint64_t)sum1 << 32) | sum0; \ +} + +#define DEF_IMOP_64(NAME, NTYPE, MTYPE) \ +static uint64_t NAME(uint64_t n, uint64_t m, uint64_t a, uint8_t p, bool neg) \ +{ \ + uint64_t sum = 0; \ + /* Apply P to N as a mask, making the inactive elements 0. */ \ + n &= expand_pred_h(p); \ + sum += (NTYPE)(n >> 0) * (MTYPE)(m >> 0); \ + sum += (NTYPE)(n >> 16) * (MTYPE)(m >> 16); \ + sum += (NTYPE)(n >> 32) * (MTYPE)(m >> 32); \ + sum += (NTYPE)(n >> 48) * (MTYPE)(m >> 48); \ + return neg ? a - sum : a + sum; \ +} + +DEF_IMOP_32(smopa_s, int8_t, int8_t) +DEF_IMOP_32(umopa_s, uint8_t, uint8_t) +DEF_IMOP_32(sumopa_s, int8_t, uint8_t) +DEF_IMOP_32(usmopa_s, uint8_t, int8_t) + +DEF_IMOP_64(smopa_d, int16_t, int16_t) +DEF_IMOP_64(umopa_d, uint16_t, uint16_t) +DEF_IMOP_64(sumopa_d, int16_t, uint16_t) +DEF_IMOP_64(usmopa_d, uint16_t, int16_t) + +#define DEF_IMOPH(NAME) \ + void HELPER(sme_##NAME)(void *vza, void *vzn, void *vzm, void *vpn, \ + void *vpm, uint32_t desc) \ + { do_imopa(vza, vzn, vzm, vpn, vpm, desc, NAME); } + +DEF_IMOPH(smopa_s) +DEF_IMOPH(umopa_s) +DEF_IMOPH(sumopa_s) +DEF_IMOPH(usmopa_s) +DEF_IMOPH(smopa_d) +DEF_IMOPH(umopa_d) +DEF_IMOPH(sumopa_d) +DEF_IMOPH(usmopa_d) diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ TRANS_FEAT(FMOPA_d, aa64_sme_f64f64, do_outprod_fpst, a, MO_64, gen_helper_sme_f /* TODO: FEAT_EBF16 */ TRANS_FEAT(BFMOPA, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_bfmopa) + +TRANS_FEAT(SMOPA_s, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_smopa_s) +TRANS_FEAT(UMOPA_s, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_umopa_s) +TRANS_FEAT(SUMOPA_s, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_sumopa_s) +TRANS_FEAT(USMOPA_s, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_usmopa_s) + +TRANS_FEAT(SMOPA_d, aa64_sme_i16i64, do_outprod, a, MO_64, gen_helper_sme_smopa_d) +TRANS_FEAT(UMOPA_d, aa64_sme_i16i64, do_outprod, a, MO_64, gen_helper_sme_umopa_d) +TRANS_FEAT(SUMOPA_d, aa64_sme_i16i64, do_outprod, a, MO_64, gen_helper_sme_sumopa_d) +TRANS_FEAT(USMOPA_d, aa64_sme_i16i64, do_outprod, a, MO_64, gen_helper_sme_usmopa_d) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This is an SVE instruction that operates using the SVE vector length but that it is present only if SME is implemented. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-29-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sve.decode | 20 +++++++++++++ target/arm/translate-sve.c | 57 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 77 insertions(+) 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 @@ BFMLALT_zzxw 01100100 11 1 ..... 0100.1 ..... ..... @rrxr_3a esz=2 ### SVE2 floating-point bfloat16 dot-product (indexed) BFDOT_zzxz 01100100 01 1 ..... 010000 ..... ..... @rrxr_2 esz=2 + +### SVE broadcast predicate element + +&psel esz pd pn pm rv imm +%psel_rv 16:2 !function=plus_12 +%psel_imm_b 22:2 19:2 +%psel_imm_h 22:2 20:1 +%psel_imm_s 22:2 +%psel_imm_d 23:1 +@psel ........ .. . ... .. .. pn:4 . pm:4 . pd:4 \ + &psel rv=%psel_rv + +PSEL 00100101 .. 1 ..1 .. 01 .... 0 .... 0 .... \ + @psel esz=0 imm=%psel_imm_b +PSEL 00100101 .. 1 .10 .. 01 .... 0 .... 0 .... \ + @psel esz=1 imm=%psel_imm_h +PSEL 00100101 .. 1 100 .. 01 .... 0 .... 0 .... \ + @psel esz=2 imm=%psel_imm_s +PSEL 00100101 .1 1 000 .. 01 .... 0 .... 0 .... \ + @psel esz=3 imm=%psel_imm_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 do_BFMLAL_zzxw(DisasContext *s, arg_rrxr_esz *a, bool sel) TRANS_FEAT(BFMLALB_zzxw, aa64_sve_bf16, do_BFMLAL_zzxw, a, false) TRANS_FEAT(BFMLALT_zzxw, aa64_sve_bf16, do_BFMLAL_zzxw, a, true) + +static bool trans_PSEL(DisasContext *s, arg_psel *a) +{ + int vl = vec_full_reg_size(s); + int pl = pred_gvec_reg_size(s); + int elements = vl >> a->esz; + TCGv_i64 tmp, didx, dbit; + TCGv_ptr ptr; + + if (!dc_isar_feature(aa64_sme, s)) { + return false; + } + if (!sve_access_check(s)) { + return true; + } + + tmp = tcg_temp_new_i64(); + dbit = tcg_temp_new_i64(); + didx = tcg_temp_new_i64(); + ptr = tcg_temp_new_ptr(); + + /* Compute the predicate element. */ + tcg_gen_addi_i64(tmp, cpu_reg(s, a->rv), a->imm); + if (is_power_of_2(elements)) { + tcg_gen_andi_i64(tmp, tmp, elements - 1); + } else { + tcg_gen_remu_i64(tmp, tmp, tcg_constant_i64(elements)); + } + + /* Extract the predicate byte and bit indices. */ + tcg_gen_shli_i64(tmp, tmp, a->esz); + tcg_gen_andi_i64(dbit, tmp, 7); + tcg_gen_shri_i64(didx, tmp, 3); + if (HOST_BIG_ENDIAN) { + tcg_gen_xori_i64(didx, didx, 7); + } + + /* Load the predicate word. */ + tcg_gen_trunc_i64_ptr(ptr, didx); + tcg_gen_add_ptr(ptr, ptr, cpu_env); + tcg_gen_ld8u_i64(tmp, ptr, pred_full_reg_offset(s, a->pm)); + + /* Extract the predicate bit and replicate to MO_64. */ + tcg_gen_shr_i64(tmp, tmp, dbit); + tcg_gen_andi_i64(tmp, tmp, 1); + tcg_gen_neg_i64(tmp, tmp); + + /* Apply to either copy the source, or write zeros. */ + tcg_gen_gvec_ands(MO_64, pred_full_reg_offset(s, a->pd), + pred_full_reg_offset(s, a->pn), tmp, pl, pl); + + tcg_temp_free_i64(tmp); + tcg_temp_free_i64(dbit); + tcg_temp_free_i64(didx); + tcg_temp_free_ptr(ptr); + return true; +} -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This is an SVE instruction that operates using the SVE vector length but that it is present only if SME is implemented. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-30-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sve.h | 2 ++ target/arm/sve.decode | 1 + target/arm/sve_helper.c | 16 ++++++++++++++++ target/arm/translate-sve.c | 2 ++ 4 files changed, 21 insertions(+) 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_revh_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sve_revw_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_revd_q, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + DEF_HELPER_FLAGS_4(sve_rbit_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sve_rbit_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sve_rbit_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) 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 @@ REVB 00000101 .. 1001 00 100 ... ..... ..... @rd_pg_rn REVH 00000101 .. 1001 01 100 ... ..... ..... @rd_pg_rn REVW 00000101 .. 1001 10 100 ... ..... ..... @rd_pg_rn RBIT 00000101 .. 1001 11 100 ... ..... ..... @rd_pg_rn +REVD 00000101 00 1011 10 100 ... ..... ..... @rd_pg_rn_e0 # SVE vector splice (predicated, destructive) SPLICE 00000101 .. 101 100 100 ... ..... ..... @rdn_pg_rm 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_ZPZ_D(sve_revh_d, uint64_t, hswap64) DO_ZPZ_D(sve_revw_d, uint64_t, wswap64) +void HELPER(sme_revd_q)(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; + + for (i = 0; i < opr_sz; i += 2) { + if (pg[H1(i)] & 1) { + uint64_t n0 = n[i + 0]; + uint64_t n1 = n[i + 1]; + d[i + 0] = n1; + d[i + 1] = n0; + } + } +} + DO_ZPZ(sve_rbit_b, uint8_t, H1, revbit8) DO_ZPZ(sve_rbit_h, uint16_t, H1_2, revbit16) DO_ZPZ(sve_rbit_s, uint32_t, H1_4, revbit32) 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 @@ TRANS_FEAT(REVH, aa64_sve, gen_gvec_ool_arg_zpz, revh_fns[a->esz], a, 0) TRANS_FEAT(REVW, aa64_sve, gen_gvec_ool_arg_zpz, a->esz == 3 ? gen_helper_sve_revw_d : NULL, a, 0) +TRANS_FEAT(REVD, aa64_sme, gen_gvec_ool_arg_zpz, gen_helper_sme_revd_q, a, 0) + TRANS_FEAT(SPLICE, aa64_sve, gen_gvec_ool_arg_zpzz, gen_helper_sve_splice, a, a->esz) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This is an SVE instruction that operates using the SVE vector length but that it is present only if SME is implemented. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-31-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 18 +++++++ target/arm/sve.decode | 5 ++ target/arm/translate-sve.c | 102 +++++++++++++++++++++++++++++++++++++ target/arm/vec_helper.c | 24 +++++++++ 4 files changed, 149 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_6(gvec_bfmlal, TCG_CALL_NO_RWG, DEF_HELPER_FLAGS_6(gvec_bfmlal_idx, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_sclamp_b, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_sclamp_h, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_sclamp_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_sclamp_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_5(gvec_uclamp_b, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_uclamp_h, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_uclamp_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_uclamp_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/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 @@ PSEL 00100101 .. 1 100 .. 01 .... 0 .... 0 .... \ @psel esz=2 imm=%psel_imm_s PSEL 00100101 .1 1 000 .. 01 .... 0 .... 0 .... \ @psel esz=3 imm=%psel_imm_d + +### SVE clamp + +SCLAMP 01000100 .. 0 ..... 110000 ..... ..... @rda_rn_rm +UCLAMP 01000100 .. 0 ..... 110001 ..... ..... @rda_rn_rm 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_PSEL(DisasContext *s, arg_psel *a) tcg_temp_free_ptr(ptr); return true; } + +static void gen_sclamp_i32(TCGv_i32 d, TCGv_i32 n, TCGv_i32 m, TCGv_i32 a) +{ + tcg_gen_smax_i32(d, a, n); + tcg_gen_smin_i32(d, d, m); +} + +static void gen_sclamp_i64(TCGv_i64 d, TCGv_i64 n, TCGv_i64 m, TCGv_i64 a) +{ + tcg_gen_smax_i64(d, a, n); + tcg_gen_smin_i64(d, d, m); +} + +static void gen_sclamp_vec(unsigned vece, TCGv_vec d, TCGv_vec n, + TCGv_vec m, TCGv_vec a) +{ + tcg_gen_smax_vec(vece, d, a, n); + tcg_gen_smin_vec(vece, d, d, m); +} + +static void gen_sclamp(unsigned vece, uint32_t d, uint32_t n, uint32_t m, + uint32_t a, uint32_t oprsz, uint32_t maxsz) +{ + static const TCGOpcode vecop[] = { + INDEX_op_smin_vec, INDEX_op_smax_vec, 0 + }; + static const GVecGen4 ops[4] = { + { .fniv = gen_sclamp_vec, + .fno = gen_helper_gvec_sclamp_b, + .opt_opc = vecop, + .vece = MO_8 }, + { .fniv = gen_sclamp_vec, + .fno = gen_helper_gvec_sclamp_h, + .opt_opc = vecop, + .vece = MO_16 }, + { .fni4 = gen_sclamp_i32, + .fniv = gen_sclamp_vec, + .fno = gen_helper_gvec_sclamp_s, + .opt_opc = vecop, + .vece = MO_32 }, + { .fni8 = gen_sclamp_i64, + .fniv = gen_sclamp_vec, + .fno = gen_helper_gvec_sclamp_d, + .opt_opc = vecop, + .vece = MO_64, + .prefer_i64 = TCG_TARGET_REG_BITS == 64 } + }; + tcg_gen_gvec_4(d, n, m, a, oprsz, maxsz, &ops[vece]); +} + +TRANS_FEAT(SCLAMP, aa64_sme, gen_gvec_fn_arg_zzzz, gen_sclamp, a) + +static void gen_uclamp_i32(TCGv_i32 d, TCGv_i32 n, TCGv_i32 m, TCGv_i32 a) +{ + tcg_gen_umax_i32(d, a, n); + tcg_gen_umin_i32(d, d, m); +} + +static void gen_uclamp_i64(TCGv_i64 d, TCGv_i64 n, TCGv_i64 m, TCGv_i64 a) +{ + tcg_gen_umax_i64(d, a, n); + tcg_gen_umin_i64(d, d, m); +} + +static void gen_uclamp_vec(unsigned vece, TCGv_vec d, TCGv_vec n, + TCGv_vec m, TCGv_vec a) +{ + tcg_gen_umax_vec(vece, d, a, n); + tcg_gen_umin_vec(vece, d, d, m); +} + +static void gen_uclamp(unsigned vece, uint32_t d, uint32_t n, uint32_t m, + uint32_t a, uint32_t oprsz, uint32_t maxsz) +{ + static const TCGOpcode vecop[] = { + INDEX_op_umin_vec, INDEX_op_umax_vec, 0 + }; + static const GVecGen4 ops[4] = { + { .fniv = gen_uclamp_vec, + .fno = gen_helper_gvec_uclamp_b, + .opt_opc = vecop, + .vece = MO_8 }, + { .fniv = gen_uclamp_vec, + .fno = gen_helper_gvec_uclamp_h, + .opt_opc = vecop, + .vece = MO_16 }, + { .fni4 = gen_uclamp_i32, + .fniv = gen_uclamp_vec, + .fno = gen_helper_gvec_uclamp_s, + .opt_opc = vecop, + .vece = MO_32 }, + { .fni8 = gen_uclamp_i64, + .fniv = gen_uclamp_vec, + .fno = gen_helper_gvec_uclamp_d, + .opt_opc = vecop, + .vece = MO_64, + .prefer_i64 = TCG_TARGET_REG_BITS == 64 } + }; + tcg_gen_gvec_4(d, n, m, a, oprsz, maxsz, &ops[vece]); +} + +TRANS_FEAT(UCLAMP, aa64_sme, gen_gvec_fn_arg_zzzz, gen_uclamp, a) 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_bfmlal_idx)(void *vd, void *vn, void *vm, } clear_tail(d, opr_sz, simd_maxsz(desc)); } + +#define DO_CLAMP(NAME, TYPE) \ +void HELPER(NAME)(void *d, void *n, void *m, void *a, uint32_t desc) \ +{ \ + intptr_t i, opr_sz = simd_oprsz(desc); \ + for (i = 0; i < opr_sz; i += sizeof(TYPE)) { \ + TYPE aa = *(TYPE *)(a + i); \ + TYPE nn = *(TYPE *)(n + i); \ + TYPE mm = *(TYPE *)(m + i); \ + TYPE dd = MIN(MAX(aa, nn), mm); \ + *(TYPE *)(d + i) = dd; \ + } \ + clear_tail(d, opr_sz, simd_maxsz(desc)); \ +} + +DO_CLAMP(gvec_sclamp_b, int8_t) +DO_CLAMP(gvec_sclamp_h, int16_t) +DO_CLAMP(gvec_sclamp_s, int32_t) +DO_CLAMP(gvec_sclamp_d, int64_t) + +DO_CLAMP(gvec_uclamp_b, uint8_t) +DO_CLAMP(gvec_uclamp_h, uint16_t) +DO_CLAMP(gvec_uclamp_s, uint32_t) +DO_CLAMP(gvec_uclamp_d, uint64_t) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> We can handle both exception entry and exception return by hooking into aarch64_sve_change_el. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-32-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 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 @@ void aarch64_sve_change_el(CPUARMState *env, int old_el, return; } + old_a64 = old_el ? arm_el_is_aa64(env, old_el) : el0_a64; + new_a64 = new_el ? arm_el_is_aa64(env, new_el) : el0_a64; + + /* + * Both AArch64.TakeException and AArch64.ExceptionReturn + * invoke ResetSVEState when taking an exception from, or + * returning to, AArch32 state when PSTATE.SM is enabled. + */ + if (old_a64 != new_a64 && FIELD_EX64(env->svcr, SVCR, SM)) { + arm_reset_sve_state(env); + 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 @@ -XXX,XX +XXX,XX @@ void aarch64_sve_change_el(CPUARMState *env, int old_el, * we already have the correct register contents when encountering the * vq0->vq0 transition between EL0->EL1. */ - old_a64 = old_el ? arm_el_is_aa64(env, old_el) : el0_a64; old_len = (old_a64 && !sve_exception_el(env, old_el) ? sve_vqm1_for_el(env, old_el) : 0); - new_a64 = new_el ? arm_el_is_aa64(env, new_el) : el0_a64; new_len = (new_a64 && !sve_exception_el(env, new_el) ? sve_vqm1_for_el(env, new_el) : 0); -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Note that SME remains effectively disabled for user-only, because we do not yet set CPACR_EL1.SMEN. This needs to wait until the kernel ABI is implemented. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-33-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/system/arm/emulation.rst | 4 ++++ target/arm/cpu64.c | 11 +++++++++++ 2 files changed, 15 insertions(+) diff --git a/docs/system/arm/emulation.rst b/docs/system/arm/emulation.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/emulation.rst +++ b/docs/system/arm/emulation.rst @@ -XXX,XX +XXX,XX @@ the following architecture extensions: - FEAT_SHA512 (Advanced SIMD SHA512 instructions) - FEAT_SM3 (Advanced SIMD SM3 instructions) - FEAT_SM4 (Advanced SIMD SM4 instructions) +- FEAT_SME (Scalable Matrix Extension) +- FEAT_SME_FA64 (Full A64 instruction set in Streaming SVE mode) +- FEAT_SME_F64F64 (Double-precision floating-point outer product instructions) +- FEAT_SME_I16I64 (16-bit to 64-bit integer widening outer product instructions) - FEAT_SPECRES (Speculation restriction instructions) - FEAT_SSBS (Speculative Store Bypass Safe) - FEAT_TLBIOS (TLB invalidate instructions in Outer Shareable domain) 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_max_initfn(Object *obj) */ t = FIELD_DP64(t, ID_AA64PFR1, MTE, 3); /* FEAT_MTE3 */ t = FIELD_DP64(t, ID_AA64PFR1, RAS_FRAC, 0); /* FEAT_RASv1p1 + FEAT_DoubleFault */ + t = FIELD_DP64(t, ID_AA64PFR1, SME, 1); /* FEAT_SME */ t = FIELD_DP64(t, ID_AA64PFR1, CSV2_FRAC, 0); /* FEAT_CSV2_2 */ cpu->isar.id_aa64pfr1 = t; @@ -XXX,XX +XXX,XX @@ static void aarch64_max_initfn(Object *obj) t = FIELD_DP64(t, ID_AA64DFR0, PMUVER, 5); /* FEAT_PMUv3p4 */ cpu->isar.id_aa64dfr0 = t; + t = cpu->isar.id_aa64smfr0; + t = FIELD_DP64(t, ID_AA64SMFR0, F32F32, 1); /* FEAT_SME */ + t = FIELD_DP64(t, ID_AA64SMFR0, B16F32, 1); /* FEAT_SME */ + t = FIELD_DP64(t, ID_AA64SMFR0, F16F32, 1); /* FEAT_SME */ + t = FIELD_DP64(t, ID_AA64SMFR0, I8I32, 0xf); /* FEAT_SME */ + t = FIELD_DP64(t, ID_AA64SMFR0, F64F64, 1); /* FEAT_SME_F64F64 */ + t = FIELD_DP64(t, ID_AA64SMFR0, I16I64, 0xf); /* FEAT_SME_I16I64 */ + t = FIELD_DP64(t, ID_AA64SMFR0, FA64, 1); /* FEAT_SME_FA64 */ + cpu->isar.id_aa64smfr0 = t; + /* Replicate the same data to the 32-bit id registers. */ aa32_max_features(cpu); -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-34-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/target_cpu.h | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/linux-user/aarch64/target_cpu.h b/linux-user/aarch64/target_cpu.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/target_cpu.h +++ b/linux-user/aarch64/target_cpu.h @@ -XXX,XX +XXX,XX @@ static inline void cpu_clone_regs_parent(CPUARMState *env, unsigned flags) static inline void cpu_set_tls(CPUARMState *env, target_ulong newtls) { - /* Note that AArch64 Linux keeps the TLS pointer in TPIDR; this is + /* + * Note that AArch64 Linux keeps the TLS pointer in TPIDR; this is * different from AArch32 Linux, which uses TPIDRRO. */ env->cp15.tpidr_el[0] = newtls; + /* TPIDR2_EL0 is cleared with CLONE_SETTLS. */ + env->cp15.tpidr2_el0 = 0; } static inline abi_ulong get_sp_from_cpustate(CPUARMState *state) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-35-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/cpu_loop.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/linux-user/aarch64/cpu_loop.c b/linux-user/aarch64/cpu_loop.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/cpu_loop.c +++ b/linux-user/aarch64/cpu_loop.c @@ -XXX,XX +XXX,XX @@ void cpu_loop(CPUARMState *env) switch (trapnr) { case EXCP_SWI: + /* + * On syscall, PSTATE.ZA is preserved, along with the ZA matrix. + * PSTATE.SM is cleared, per SMSTOP, which does ResetSVEState. + */ + if (FIELD_EX64(env->svcr, SVCR, SM)) { + env->svcr = FIELD_DP64(env->svcr, SVCR, SM, 0); + arm_rebuild_hflags(env); + arm_reset_sve_state(env); + } ret = do_syscall(env, env->xregs[8], env->xregs[0], -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Make sure to zero the currently reserved fields. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-36-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/signal.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/linux-user/aarch64/signal.c b/linux-user/aarch64/signal.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/signal.c +++ b/linux-user/aarch64/signal.c @@ -XXX,XX +XXX,XX @@ struct target_extra_context { struct target_sve_context { struct target_aarch64_ctx head; uint16_t vl; - uint16_t reserved[3]; + uint16_t flags; + uint16_t reserved[2]; /* The actual SVE data immediately follows. It is laid out * according to TARGET_SVE_SIG_{Z,P}REG_OFFSET, based off of * the original struct pointer. @@ -XXX,XX +XXX,XX @@ struct target_sve_context { #define TARGET_SVE_SIG_CONTEXT_SIZE(VQ) \ (TARGET_SVE_SIG_PREG_OFFSET(VQ, 17)) +#define TARGET_SVE_SIG_FLAG_SM 1 + struct target_rt_sigframe { struct target_siginfo info; struct target_ucontext uc; @@ -XXX,XX +XXX,XX @@ static void target_setup_sve_record(struct target_sve_context *sve, { int i, j; + memset(sve, 0, sizeof(*sve)); __put_user(TARGET_SVE_MAGIC, &sve->head.magic); __put_user(size, &sve->head.size); __put_user(vq * TARGET_SVE_VQ_BYTES, &sve->vl); + if (FIELD_EX64(env->svcr, SVCR, SM)) { + __put_user(TARGET_SVE_SIG_FLAG_SM, &sve->flags); + } /* Note that SVE regs are stored as a byte stream, with each byte element * at a subsequent address. This corresponds to a little-endian store -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Fold the return value setting into the goto, so each point of failure need not do both. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-37-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/signal.c | 26 +++++++++++--------------- 1 file changed, 11 insertions(+), 15 deletions(-) diff --git a/linux-user/aarch64/signal.c b/linux-user/aarch64/signal.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/signal.c +++ b/linux-user/aarch64/signal.c @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, struct target_sve_context *sve = NULL; uint64_t extra_datap = 0; bool used_extra = false; - bool err = false; int vq = 0, sve_size = 0; target_restore_general_frame(env, sf); @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, switch (magic) { case 0: if (size != 0) { - err = true; - goto exit; + goto err; } if (used_extra) { ctx = NULL; @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, case TARGET_FPSIMD_MAGIC: if (fpsimd || size != sizeof(struct target_fpsimd_context)) { - err = true; - goto exit; + goto err; } fpsimd = (struct target_fpsimd_context *)ctx; break; @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, break; } } - err = true; - goto exit; + goto err; case TARGET_EXTRA_MAGIC: if (extra || size != sizeof(struct target_extra_context)) { - err = true; - goto exit; + goto err; } __get_user(extra_datap, &((struct target_extra_context *)ctx)->datap); @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, /* Unknown record -- we certainly didn't generate it. * Did we in fact get out of sync? */ - err = true; - goto exit; + goto err; } ctx = (void *)ctx + size; } @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, if (fpsimd) { target_restore_fpsimd_record(env, fpsimd); } else { - err = true; + goto err; } /* SVE data, if present, overwrites FPSIMD data. */ if (sve) { target_restore_sve_record(env, sve, vq); } - - exit: unlock_user(extra, extra_datap, 0); - return err; + return 0; + + err: + unlock_user(extra, extra_datap, 0); + return 1; } static abi_ulong get_sigframe(struct target_sigaction *ka, -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> In parse_user_sigframe, the kernel rejects duplicate sve records, or records that are smaller than the header. We were silently allowing these cases to pass, dropping the record. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-38-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/signal.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/linux-user/aarch64/signal.c b/linux-user/aarch64/signal.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/signal.c +++ b/linux-user/aarch64/signal.c @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, break; case TARGET_SVE_MAGIC: + if (sve || size < sizeof(struct target_sve_context)) { + goto err; + } if (cpu_isar_feature(aa64_sve, env_archcpu(env))) { vq = sve_vq(env); sve_size = QEMU_ALIGN_UP(TARGET_SVE_SIG_CONTEXT_SIZE(vq), 16); - if (!sve && size == sve_size) { + if (size == sve_size) { sve = (struct target_sve_context *)ctx; break; } -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-39-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/signal.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/linux-user/aarch64/signal.c b/linux-user/aarch64/signal.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/signal.c +++ b/linux-user/aarch64/signal.c @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, __get_user(extra_size, &((struct target_extra_context *)ctx)->size); extra = lock_user(VERIFY_READ, extra_datap, extra_size, 0); + if (!extra) { + return 1; + } break; default: -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Move the checks out of the parsing loop and into the restore function. This more closely mirrors the code structure in the kernel, and is slightly clearer. Reject rather than silently skip incorrect VL and SVE record sizes, bringing our checks in to line with those the kernel does. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-40-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/signal.c | 51 +++++++++++++++++++++++++------------ 1 file changed, 35 insertions(+), 16 deletions(-) diff --git a/linux-user/aarch64/signal.c b/linux-user/aarch64/signal.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/signal.c +++ b/linux-user/aarch64/signal.c @@ -XXX,XX +XXX,XX @@ static void target_restore_fpsimd_record(CPUARMState *env, } } -static void target_restore_sve_record(CPUARMState *env, - struct target_sve_context *sve, int vq) +static bool target_restore_sve_record(CPUARMState *env, + struct target_sve_context *sve, + int size) { - int i, j; + int i, j, vl, vq; - /* Note that SVE regs are stored as a byte stream, with each byte element + if (!cpu_isar_feature(aa64_sve, env_archcpu(env))) { + return false; + } + + __get_user(vl, &sve->vl); + vq = sve_vq(env); + + /* Reject mismatched VL. */ + if (vl != vq * TARGET_SVE_VQ_BYTES) { + return false; + } + + /* Accept empty record -- used to clear PSTATE.SM. */ + if (size <= sizeof(*sve)) { + return true; + } + + /* Reject non-empty but incomplete record. */ + if (size < TARGET_SVE_SIG_CONTEXT_SIZE(vq)) { + return false; + } + + /* + * Note that SVE regs are stored as a byte stream, with each byte element * at a subsequent address. This corresponds to a little-endian load * of our 64-bit hunks. */ @@ -XXX,XX +XXX,XX @@ static void target_restore_sve_record(CPUARMState *env, } } } + return true; } static int target_restore_sigframe(CPUARMState *env, @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, struct target_sve_context *sve = NULL; uint64_t extra_datap = 0; bool used_extra = false; - int vq = 0, sve_size = 0; + int sve_size = 0; target_restore_general_frame(env, sf); @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, if (sve || size < sizeof(struct target_sve_context)) { goto err; } - if (cpu_isar_feature(aa64_sve, env_archcpu(env))) { - vq = sve_vq(env); - sve_size = QEMU_ALIGN_UP(TARGET_SVE_SIG_CONTEXT_SIZE(vq), 16); - if (size == sve_size) { - sve = (struct target_sve_context *)ctx; - break; - } - } - goto err; + sve = (struct target_sve_context *)ctx; + sve_size = size; + break; case TARGET_EXTRA_MAGIC: if (extra || size != sizeof(struct target_extra_context)) { @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, } /* SVE data, if present, overwrites FPSIMD data. */ - if (sve) { - target_restore_sve_record(env, sve, vq); + if (sve && !target_restore_sve_record(env, sve, sve_size)) { + goto err; } unlock_user(extra, extra_datap, 0); return 0; -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Set the SM bit in the SVE record on signal delivery, create the ZA record. Restore SM and ZA state according to the records present on return. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-41-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/signal.c | 167 +++++++++++++++++++++++++++++++++--- 1 file changed, 154 insertions(+), 13 deletions(-) diff --git a/linux-user/aarch64/signal.c b/linux-user/aarch64/signal.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/signal.c +++ b/linux-user/aarch64/signal.c @@ -XXX,XX +XXX,XX @@ struct target_sve_context { #define TARGET_SVE_SIG_FLAG_SM 1 +#define TARGET_ZA_MAGIC 0x54366345 + +struct target_za_context { + struct target_aarch64_ctx head; + uint16_t vl; + uint16_t reserved[3]; + /* The actual ZA data immediately follows. */ +}; + +#define TARGET_ZA_SIG_REGS_OFFSET \ + QEMU_ALIGN_UP(sizeof(struct target_za_context), TARGET_SVE_VQ_BYTES) +#define TARGET_ZA_SIG_ZAV_OFFSET(VQ, N) \ + (TARGET_ZA_SIG_REGS_OFFSET + (VQ) * TARGET_SVE_VQ_BYTES * (N)) +#define TARGET_ZA_SIG_CONTEXT_SIZE(VQ) \ + TARGET_ZA_SIG_ZAV_OFFSET(VQ, VQ * TARGET_SVE_VQ_BYTES) + struct target_rt_sigframe { struct target_siginfo info; struct target_ucontext uc; @@ -XXX,XX +XXX,XX @@ static void target_setup_end_record(struct target_aarch64_ctx *end) } static void target_setup_sve_record(struct target_sve_context *sve, - CPUARMState *env, int vq, int size) + CPUARMState *env, int size) { - int i, j; + int i, j, vq = sve_vq(env); memset(sve, 0, sizeof(*sve)); __put_user(TARGET_SVE_MAGIC, &sve->head.magic); @@ -XXX,XX +XXX,XX @@ static void target_setup_sve_record(struct target_sve_context *sve, } } +static void target_setup_za_record(struct target_za_context *za, + CPUARMState *env, int size) +{ + int vq = sme_vq(env); + int vl = vq * TARGET_SVE_VQ_BYTES; + int i, j; + + memset(za, 0, sizeof(*za)); + __put_user(TARGET_ZA_MAGIC, &za->head.magic); + __put_user(size, &za->head.size); + __put_user(vl, &za->vl); + + if (size == TARGET_ZA_SIG_CONTEXT_SIZE(0)) { + return; + } + assert(size == TARGET_ZA_SIG_CONTEXT_SIZE(vq)); + + /* + * Note that ZA vectors are stored as a byte stream, + * with each byte element at a subsequent address. + */ + for (i = 0; i < vl; ++i) { + uint64_t *z = (void *)za + TARGET_ZA_SIG_ZAV_OFFSET(vq, i); + for (j = 0; j < vq * 2; ++j) { + __put_user_e(env->zarray[i].d[j], z + j, le); + } + } +} + static void target_restore_general_frame(CPUARMState *env, struct target_rt_sigframe *sf) { @@ -XXX,XX +XXX,XX @@ static void target_restore_fpsimd_record(CPUARMState *env, static bool target_restore_sve_record(CPUARMState *env, struct target_sve_context *sve, - int size) + int size, int *svcr) { - int i, j, vl, vq; + int i, j, vl, vq, flags; + bool sm; - if (!cpu_isar_feature(aa64_sve, env_archcpu(env))) { + __get_user(vl, &sve->vl); + __get_user(flags, &sve->flags); + + sm = flags & TARGET_SVE_SIG_FLAG_SM; + + /* The cpu must support Streaming or Non-streaming SVE. */ + if (sm + ? !cpu_isar_feature(aa64_sme, env_archcpu(env)) + : !cpu_isar_feature(aa64_sve, env_archcpu(env))) { return false; } - __get_user(vl, &sve->vl); - vq = sve_vq(env); + /* + * Note that we cannot use sve_vq() because that depends on the + * current setting of PSTATE.SM, not the state to be restored. + */ + vq = sve_vqm1_for_el_sm(env, 0, sm) + 1; /* Reject mismatched VL. */ if (vl != vq * TARGET_SVE_VQ_BYTES) { @@ -XXX,XX +XXX,XX @@ static bool target_restore_sve_record(CPUARMState *env, return false; } + *svcr = FIELD_DP64(*svcr, SVCR, SM, sm); + /* * Note that SVE regs are stored as a byte stream, with each byte element * at a subsequent address. This corresponds to a little-endian load @@ -XXX,XX +XXX,XX @@ static bool target_restore_sve_record(CPUARMState *env, return true; } +static bool target_restore_za_record(CPUARMState *env, + struct target_za_context *za, + int size, int *svcr) +{ + int i, j, vl, vq; + + if (!cpu_isar_feature(aa64_sme, env_archcpu(env))) { + return false; + } + + __get_user(vl, &za->vl); + vq = sme_vq(env); + + /* Reject mismatched VL. */ + if (vl != vq * TARGET_SVE_VQ_BYTES) { + return false; + } + + /* Accept empty record -- used to clear PSTATE.ZA. */ + if (size <= TARGET_ZA_SIG_CONTEXT_SIZE(0)) { + return true; + } + + /* Reject non-empty but incomplete record. */ + if (size < TARGET_ZA_SIG_CONTEXT_SIZE(vq)) { + return false; + } + + *svcr = FIELD_DP64(*svcr, SVCR, ZA, 1); + + for (i = 0; i < vl; ++i) { + uint64_t *z = (void *)za + TARGET_ZA_SIG_ZAV_OFFSET(vq, i); + for (j = 0; j < vq * 2; ++j) { + __get_user_e(env->zarray[i].d[j], z + j, le); + } + } + return true; +} + static int target_restore_sigframe(CPUARMState *env, struct target_rt_sigframe *sf) { struct target_aarch64_ctx *ctx, *extra = NULL; struct target_fpsimd_context *fpsimd = NULL; struct target_sve_context *sve = NULL; + struct target_za_context *za = NULL; uint64_t extra_datap = 0; bool used_extra = false; int sve_size = 0; + int za_size = 0; + int svcr = 0; target_restore_general_frame(env, sf); @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, sve_size = size; break; + case TARGET_ZA_MAGIC: + if (za || size < sizeof(struct target_za_context)) { + goto err; + } + za = (struct target_za_context *)ctx; + za_size = size; + break; + case TARGET_EXTRA_MAGIC: if (extra || size != sizeof(struct target_extra_context)) { goto err; @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, } /* SVE data, if present, overwrites FPSIMD data. */ - if (sve && !target_restore_sve_record(env, sve, sve_size)) { + if (sve && !target_restore_sve_record(env, sve, sve_size, &svcr)) { goto err; } + if (za && !target_restore_za_record(env, za, za_size, &svcr)) { + goto err; + } + if (env->svcr != svcr) { + env->svcr = svcr; + arm_rebuild_hflags(env); + } unlock_user(extra, extra_datap, 0); return 0; @@ -XXX,XX +XXX,XX @@ static void target_setup_frame(int usig, struct target_sigaction *ka, .total_size = offsetof(struct target_rt_sigframe, uc.tuc_mcontext.__reserved), }; - int fpsimd_ofs, fr_ofs, sve_ofs = 0, vq = 0, sve_size = 0; + int fpsimd_ofs, fr_ofs, sve_ofs = 0, za_ofs = 0; + int sve_size = 0, za_size = 0; struct target_rt_sigframe *frame; struct target_rt_frame_record *fr; abi_ulong frame_addr, return_addr; @@ -XXX,XX +XXX,XX @@ static void target_setup_frame(int usig, struct target_sigaction *ka, &layout); /* SVE state needs saving only if it exists. */ - if (cpu_isar_feature(aa64_sve, env_archcpu(env))) { - vq = sve_vq(env); - sve_size = QEMU_ALIGN_UP(TARGET_SVE_SIG_CONTEXT_SIZE(vq), 16); + if (cpu_isar_feature(aa64_sve, env_archcpu(env)) || + cpu_isar_feature(aa64_sme, env_archcpu(env))) { + sve_size = QEMU_ALIGN_UP(TARGET_SVE_SIG_CONTEXT_SIZE(sve_vq(env)), 16); sve_ofs = alloc_sigframe_space(sve_size, &layout); } + if (cpu_isar_feature(aa64_sme, env_archcpu(env))) { + /* ZA state needs saving only if it is enabled. */ + if (FIELD_EX64(env->svcr, SVCR, ZA)) { + za_size = TARGET_ZA_SIG_CONTEXT_SIZE(sme_vq(env)); + } else { + za_size = TARGET_ZA_SIG_CONTEXT_SIZE(0); + } + za_ofs = alloc_sigframe_space(za_size, &layout); + } if (layout.extra_ofs) { /* Reserve space for the extra end marker. The standard end marker @@ -XXX,XX +XXX,XX @@ static void target_setup_frame(int usig, struct target_sigaction *ka, target_setup_end_record((void *)frame + layout.extra_end_ofs); } if (sve_ofs) { - target_setup_sve_record((void *)frame + sve_ofs, env, vq, sve_size); + target_setup_sve_record((void *)frame + sve_ofs, env, sve_size); + } + if (za_ofs) { + target_setup_za_record((void *)frame + za_ofs, env, za_size); } /* Set up the stack frame for unwinding. */ @@ -XXX,XX +XXX,XX @@ static void target_setup_frame(int usig, struct target_sigaction *ka, env->btype = 2; } + /* + * Invoke the signal handler with both SM and ZA disabled. + * When clearing SM, ResetSVEState, per SMSTOP. + */ + if (FIELD_EX64(env->svcr, SVCR, SM)) { + arm_reset_sve_state(env); + } + if (env->svcr) { + env->svcr = 0; + arm_rebuild_hflags(env); + } + if (info) { tswap_siginfo(&frame->info, info); env->xregs[1] = frame_addr + offsetof(struct target_rt_sigframe, info); -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Add "sve" to the sve prctl functions, to distinguish them from the coming "sme" prctls with similar names. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-42-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/target_prctl.h | 8 ++++---- linux-user/syscall.c | 12 ++++++------ 2 files changed, 10 insertions(+), 10 deletions(-) diff --git a/linux-user/aarch64/target_prctl.h b/linux-user/aarch64/target_prctl.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/target_prctl.h +++ b/linux-user/aarch64/target_prctl.h @@ -XXX,XX +XXX,XX @@ #ifndef AARCH64_TARGET_PRCTL_H #define AARCH64_TARGET_PRCTL_H -static abi_long do_prctl_get_vl(CPUArchState *env) +static abi_long do_prctl_sve_get_vl(CPUArchState *env) { ARMCPU *cpu = env_archcpu(env); if (cpu_isar_feature(aa64_sve, cpu)) { @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl_get_vl(CPUArchState *env) } return -TARGET_EINVAL; } -#define do_prctl_get_vl do_prctl_get_vl +#define do_prctl_sve_get_vl do_prctl_sve_get_vl -static abi_long do_prctl_set_vl(CPUArchState *env, abi_long arg2) +static abi_long do_prctl_sve_set_vl(CPUArchState *env, abi_long arg2) { /* * We cannot support either PR_SVE_SET_VL_ONEXEC or PR_SVE_VL_INHERIT. @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl_set_vl(CPUArchState *env, abi_long arg2) } return -TARGET_EINVAL; } -#define do_prctl_set_vl do_prctl_set_vl +#define do_prctl_sve_set_vl do_prctl_sve_set_vl static abi_long do_prctl_reset_keys(CPUArchState *env, abi_long arg2) { diff --git a/linux-user/syscall.c b/linux-user/syscall.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/syscall.c +++ b/linux-user/syscall.c @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl_inval1(CPUArchState *env, abi_long arg2) #ifndef do_prctl_set_fp_mode #define do_prctl_set_fp_mode do_prctl_inval1 #endif -#ifndef do_prctl_get_vl -#define do_prctl_get_vl do_prctl_inval0 +#ifndef do_prctl_sve_get_vl +#define do_prctl_sve_get_vl do_prctl_inval0 #endif -#ifndef do_prctl_set_vl -#define do_prctl_set_vl do_prctl_inval1 +#ifndef do_prctl_sve_set_vl +#define do_prctl_sve_set_vl do_prctl_inval1 #endif #ifndef do_prctl_reset_keys #define do_prctl_reset_keys do_prctl_inval1 @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl(CPUArchState *env, abi_long option, abi_long arg2, case PR_SET_FP_MODE: return do_prctl_set_fp_mode(env, arg2); case PR_SVE_GET_VL: - return do_prctl_get_vl(env); + return do_prctl_sve_get_vl(env); case PR_SVE_SET_VL: - return do_prctl_set_vl(env, arg2); + return do_prctl_sve_set_vl(env, arg2); case PR_PAC_RESET_KEYS: if (arg3 || arg4 || arg5) { return -TARGET_EINVAL; -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> These prctl set the Streaming SVE vector length, which may be completely different from the Normal SVE vector length. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-43-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/target_prctl.h | 54 +++++++++++++++++++++++++++++++ linux-user/syscall.c | 16 +++++++++ 2 files changed, 70 insertions(+) diff --git a/linux-user/aarch64/target_prctl.h b/linux-user/aarch64/target_prctl.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/target_prctl.h +++ b/linux-user/aarch64/target_prctl.h @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl_sve_get_vl(CPUArchState *env) { ARMCPU *cpu = env_archcpu(env); if (cpu_isar_feature(aa64_sve, cpu)) { + /* PSTATE.SM is always unset on syscall entry. */ return sve_vq(env) * 16; } return -TARGET_EINVAL; @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl_sve_set_vl(CPUArchState *env, abi_long arg2) && arg2 >= 0 && arg2 <= 512 * 16 && !(arg2 & 15)) { uint32_t vq, old_vq; + /* PSTATE.SM is always unset on syscall entry. */ old_vq = sve_vq(env); /* @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl_sve_set_vl(CPUArchState *env, abi_long arg2) } #define do_prctl_sve_set_vl do_prctl_sve_set_vl +static abi_long do_prctl_sme_get_vl(CPUArchState *env) +{ + ARMCPU *cpu = env_archcpu(env); + if (cpu_isar_feature(aa64_sme, cpu)) { + return sme_vq(env) * 16; + } + return -TARGET_EINVAL; +} +#define do_prctl_sme_get_vl do_prctl_sme_get_vl + +static abi_long do_prctl_sme_set_vl(CPUArchState *env, abi_long arg2) +{ + /* + * We cannot support either PR_SME_SET_VL_ONEXEC or PR_SME_VL_INHERIT. + * Note the kernel definition of sve_vl_valid allows for VQ=512, + * i.e. VL=8192, even though the architectural maximum is VQ=16. + */ + if (cpu_isar_feature(aa64_sme, env_archcpu(env)) + && arg2 >= 0 && arg2 <= 512 * 16 && !(arg2 & 15)) { + int vq, old_vq; + + old_vq = sme_vq(env); + + /* + * Bound the value of vq, so that we know that it fits into + * the 4-bit field in SMCR_EL1. Because PSTATE.SM is cleared + * on syscall entry, we are not modifying the current SVE + * vector length. + */ + vq = MAX(arg2 / 16, 1); + vq = MIN(vq, 16); + env->vfp.smcr_el[1] = + FIELD_DP64(env->vfp.smcr_el[1], SMCR, LEN, vq - 1); + + /* Delay rebuilding hflags until we know if ZA must change. */ + vq = sve_vqm1_for_el_sm(env, 0, true) + 1; + + if (vq != old_vq) { + /* + * PSTATE.ZA state is cleared on any change to SVL. + * We need not call arm_rebuild_hflags because PSTATE.SM was + * cleared on syscall entry, so this hasn't changed VL. + */ + env->svcr = FIELD_DP64(env->svcr, SVCR, ZA, 0); + arm_rebuild_hflags(env); + } + return vq * 16; + } + return -TARGET_EINVAL; +} +#define do_prctl_sme_set_vl do_prctl_sme_set_vl + static abi_long do_prctl_reset_keys(CPUArchState *env, abi_long arg2) { ARMCPU *cpu = env_archcpu(env); diff --git a/linux-user/syscall.c b/linux-user/syscall.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/syscall.c +++ b/linux-user/syscall.c @@ -XXX,XX +XXX,XX @@ abi_long do_arch_prctl(CPUX86State *env, int code, abi_ulong addr) #ifndef PR_SET_SYSCALL_USER_DISPATCH # define PR_SET_SYSCALL_USER_DISPATCH 59 #endif +#ifndef PR_SME_SET_VL +# define PR_SME_SET_VL 63 +# define PR_SME_GET_VL 64 +# define PR_SME_VL_LEN_MASK 0xffff +# define PR_SME_VL_INHERIT (1 << 17) +#endif #include "target_prctl.h" @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl_inval1(CPUArchState *env, abi_long arg2) #ifndef do_prctl_set_unalign #define do_prctl_set_unalign do_prctl_inval1 #endif +#ifndef do_prctl_sme_get_vl +#define do_prctl_sme_get_vl do_prctl_inval0 +#endif +#ifndef do_prctl_sme_set_vl +#define do_prctl_sme_set_vl do_prctl_inval1 +#endif static abi_long do_prctl(CPUArchState *env, abi_long option, abi_long arg2, abi_long arg3, abi_long arg4, abi_long arg5) @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl(CPUArchState *env, abi_long option, abi_long arg2, return do_prctl_sve_get_vl(env); case PR_SVE_SET_VL: return do_prctl_sve_set_vl(env, arg2); + case PR_SME_GET_VL: + return do_prctl_sme_get_vl(env); + case PR_SME_SET_VL: + return do_prctl_sme_set_vl(env, arg2); case PR_PAC_RESET_KEYS: if (arg3 || arg4 || arg5) { return -TARGET_EINVAL; -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> There's no reason to set CPACR_EL1.ZEN if SVE disabled. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-44-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ static void arm_cpu_reset(DeviceState *dev) /* and to the FP/Neon instructions */ env->cp15.cpacr_el1 = FIELD_DP64(env->cp15.cpacr_el1, CPACR_EL1, FPEN, 3); - /* and to the SVE instructions */ - env->cp15.cpacr_el1 = FIELD_DP64(env->cp15.cpacr_el1, - CPACR_EL1, ZEN, 3); - /* with reasonable vector length */ + /* and to the SVE instructions, with default vector length */ if (cpu_isar_feature(aa64_sve, cpu)) { + env->cp15.cpacr_el1 = FIELD_DP64(env->cp15.cpacr_el1, + CPACR_EL1, ZEN, 3); env->vfp.zcr_el[1] = cpu->sve_default_vq - 1; } /* -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Enable SME, TPIDR2_EL0, and FA64 if supported by the cpu. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-45-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ static void arm_cpu_reset(DeviceState *dev) CPACR_EL1, ZEN, 3); env->vfp.zcr_el[1] = cpu->sve_default_vq - 1; } + /* and for SME instructions, with default vector length, and TPIDR2 */ + if (cpu_isar_feature(aa64_sme, cpu)) { + env->cp15.sctlr_el[1] |= SCTLR_EnTP2; + env->cp15.cpacr_el1 = FIELD_DP64(env->cp15.cpacr_el1, + CPACR_EL1, SMEN, 3); + env->vfp.smcr_el[1] = cpu->sme_default_vq - 1; + if (cpu_isar_feature(aa64_sme_fa64, cpu)) { + env->vfp.smcr_el[1] = FIELD_DP64(env->vfp.smcr_el[1], + SMCR, FA64, 1); + } + } /* * Enable 48-bit address space (TODO: take reserved_va into account). * Enable TBI0 but not TBI1. -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-46-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/elfload.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/linux-user/elfload.c b/linux-user/elfload.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/elfload.c +++ b/linux-user/elfload.c @@ -XXX,XX +XXX,XX @@ enum { ARM_HWCAP2_A64_RNG = 1 << 16, ARM_HWCAP2_A64_BTI = 1 << 17, ARM_HWCAP2_A64_MTE = 1 << 18, + ARM_HWCAP2_A64_ECV = 1 << 19, + ARM_HWCAP2_A64_AFP = 1 << 20, + ARM_HWCAP2_A64_RPRES = 1 << 21, + ARM_HWCAP2_A64_MTE3 = 1 << 22, + ARM_HWCAP2_A64_SME = 1 << 23, + ARM_HWCAP2_A64_SME_I16I64 = 1 << 24, + ARM_HWCAP2_A64_SME_F64F64 = 1 << 25, + ARM_HWCAP2_A64_SME_I8I32 = 1 << 26, + ARM_HWCAP2_A64_SME_F16F32 = 1 << 27, + ARM_HWCAP2_A64_SME_B16F32 = 1 << 28, + ARM_HWCAP2_A64_SME_F32F32 = 1 << 29, + ARM_HWCAP2_A64_SME_FA64 = 1 << 30, }; #define ELF_HWCAP get_elf_hwcap() @@ -XXX,XX +XXX,XX @@ static uint32_t get_elf_hwcap2(void) GET_FEATURE_ID(aa64_rndr, ARM_HWCAP2_A64_RNG); GET_FEATURE_ID(aa64_bti, ARM_HWCAP2_A64_BTI); GET_FEATURE_ID(aa64_mte, ARM_HWCAP2_A64_MTE); + GET_FEATURE_ID(aa64_sme, (ARM_HWCAP2_A64_SME | + ARM_HWCAP2_A64_SME_F32F32 | + ARM_HWCAP2_A64_SME_B16F32 | + ARM_HWCAP2_A64_SME_F16F32 | + ARM_HWCAP2_A64_SME_I8I32)); + GET_FEATURE_ID(aa64_sme_f64f64, ARM_HWCAP2_A64_SME_F64F64); + GET_FEATURE_ID(aa64_sme_i16i64, ARM_HWCAP2_A64_SME_I16I64); + GET_FEATURE_ID(aa64_sme_fa64, ARM_HWCAP2_A64_SME_FA64); return hwcaps; } -- 2.25.1