Most of this is the Neon decodetree patches, followed by Edgar's versal cleanups. thanks -- PMM The following changes since commit 2ef486e76d64436be90f7359a3071fb2a56ce835: Merge remote-tracking branch 'remotes/marcel/tags/rdma-pull-request' into staging (2020-05-03 14:12:56 +0100) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20200504 for you to fetch changes up to 9aefc6cf9b73f66062d2f914a0136756e7a28211: target/arm: Move gen_ function typedefs to translate.h (2020-05-04 12:59:26 +0100) ---------------------------------------------------------------- target-arm queue: * Start of conversion of Neon insns to decodetree * versal board: support SD and RTC * Implement ARMv8.2-TTS2UXN * Make VQDMULL undefined when U=1 * Some minor code cleanups ---------------------------------------------------------------- Edgar E. Iglesias (11): hw/arm: versal: Remove inclusion of arm_gicv3_common.h hw/arm: versal: Move misplaced comment hw/arm: versal-virt: Fix typo xlnx-ve -> xlnx-versal hw/arm: versal: Embed the UARTs into the SoC type hw/arm: versal: Embed the GEMs into the SoC type hw/arm: versal: Embed the ADMAs into the SoC type hw/arm: versal: Embed the APUs into the SoC type hw/arm: versal: Add support for SD hw/arm: versal: Add support for the RTC hw/arm: versal-virt: Add support for SD hw/arm: versal-virt: Add support for the RTC Fredrik Strupe (1): target/arm: Make VQDMULL undefined when U=1 Peter Maydell (25): target/arm: Don't use a TLB for ARMMMUIdx_Stage2 target/arm: Use enum constant in get_phys_addr_lpae() call target/arm: Add new 's1_is_el0' argument to get_phys_addr_lpae() target/arm: Implement ARMv8.2-TTS2UXN target/arm: Use correct variable for setting 'max' cpu's ID_AA64DFR0 target/arm/translate-vfp.inc.c: Remove duplicate simd_r32 check target/arm: Don't allow Thumb Neon insns without FEATURE_NEON target/arm: Add stubs for AArch32 Neon decodetree target/arm: Convert VCMLA (vector) to decodetree target/arm: Convert VCADD (vector) to decodetree target/arm: Convert V[US]DOT (vector) to decodetree target/arm: Convert VFM[AS]L (vector) to decodetree target/arm: Convert VCMLA (scalar) to decodetree target/arm: Convert V[US]DOT (scalar) to decodetree target/arm: Convert VFM[AS]L (scalar) to decodetree target/arm: Convert Neon load/store multiple structures to decodetree target/arm: Convert Neon 'load single structure to all lanes' to decodetree target/arm: Convert Neon 'load/store single structure' to decodetree target/arm: Convert Neon 3-reg-same VADD/VSUB to decodetree target/arm: Convert Neon 3-reg-same logic ops to decodetree target/arm: Convert Neon 3-reg-same VMAX/VMIN to decodetree target/arm: Convert Neon 3-reg-same comparisons to decodetree target/arm: Convert Neon 3-reg-same VQADD/VQSUB to decodetree target/arm: Convert Neon 3-reg-same VMUL, VMLA, VMLS, VSHL to decodetree target/arm: Move gen_ function typedefs to translate.h Philippe Mathieu-Daudé (2): hw/arm/mps2-tz: Use TYPE_IOTKIT instead of hardcoded string target/arm: Use uint64_t for midr field in CPU state struct include/hw/arm/xlnx-versal.h | 31 +- target/arm/cpu-param.h | 2 +- target/arm/cpu.h | 38 ++- target/arm/translate-a64.h | 9 - target/arm/translate.h | 26 ++ target/arm/neon-dp.decode | 86 +++++ target/arm/neon-ls.decode | 52 +++ target/arm/neon-shared.decode | 66 ++++ hw/arm/mps2-tz.c | 2 +- hw/arm/xlnx-versal-virt.c | 74 ++++- hw/arm/xlnx-versal.c | 115 +++++-- target/arm/cpu.c | 3 +- target/arm/cpu64.c | 8 +- target/arm/helper.c | 183 ++++------ target/arm/translate-a64.c | 17 - target/arm/translate-neon.inc.c | 714 +++++++++++++++++++++++++++++++++++++++ target/arm/translate-vfp.inc.c | 6 - target/arm/translate.c | 716 +++------------------------------------- target/arm/Makefile.objs | 18 + 19 files changed, 1302 insertions(+), 864 deletions(-) create mode 100644 target/arm/neon-dp.decode create mode 100644 target/arm/neon-ls.decode create mode 100644 target/arm/neon-shared.decode create mode 100644 target/arm/translate-neon.inc.c From: Fredrik Strupe <fredrik@strupe.net> According to Arm ARM, VQDMULL is only valid when U=0, while having U=1 is unallocated. Signed-off-by: Fredrik Strupe <fredrik@strupe.net> Fixes: 695272dcb976 ("target-arm: Handle UNDEF cases for Neon 3-regs-different-widths") Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) {0, 0, 0, 0}, /* VMLSL */ {0, 0, 0, 9}, /* VQDMLSL */ {0, 0, 0, 0}, /* Integer VMULL */ - {0, 0, 0, 1}, /* VQDMULL */ + {0, 0, 0, 9}, /* VQDMULL */ {0, 0, 0, 0xa}, /* Polynomial VMULL */ {0, 0, 0, 7}, /* Reserved: always UNDEF */ }; -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> By using the TYPE_* definitions for devices, we can: - quickly find where devices are used with 'git-grep' - easily rename a device (one-line change). Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200428154650.21991-1-f4bug@amsat.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/mps2-tz.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/arm/mps2-tz.c b/hw/arm/mps2-tz.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/mps2-tz.c +++ b/hw/arm/mps2-tz.c @@ -XXX,XX +XXX,XX @@ static void mps2tz_common_init(MachineState *machine) exit(EXIT_FAILURE); } - sysbus_init_child_obj(OBJECT(machine), "iotkit", &mms->iotkit, + sysbus_init_child_obj(OBJECT(machine), TYPE_IOTKIT, &mms->iotkit, sizeof(mms->iotkit), mmc->armsse_type); iotkitdev = DEVICE(&mms->iotkit); object_property_set_link(OBJECT(&mms->iotkit), OBJECT(system_memory), -- 2.20.1 We define ARMMMUIdx_Stage2 as being an MMU index which uses a QEMU TLB. However we never actually use the TLB -- all stage 2 lookups are done by direct calls to get_phys_addr_lpae() followed by a physical address load via address_space_ld*(). Remove Stage2 from the list of ARM MMU indexes which correspond to real core MMU indexes, and instead put it in the set of "NOTLB" ARM MMU indexes. This allows us to drop NB_MMU_MODES to 11. It also means we can safely add support for the ARMv8.3-TTS2UXN extension, which adds permission bits to the stage 2 descriptors which define execute permission separatel for EL0 and EL1; supporting that while keeping Stage2 in a QEMU TLB would require us to use separate TLBs for "Stage2 for an EL0 access" and "Stage2 for an EL1 access", which is a lot of extra complication given we aren't even using the QEMU TLB. In the process of updating the comment on our MMU index use, fix a couple of other minor errors: * NS EL2 EL2&0 was missing from the list in the comment * some text hadn't been updated from when we bumped NB_MMU_MODES above 8 Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200330210400.11724-2-peter.maydell@linaro.org --- target/arm/cpu-param.h | 2 +- target/arm/cpu.h | 21 +++++--- target/arm/helper.c | 112 ++++------------------------------------- 3 files changed, 27 insertions(+), 108 deletions(-) diff --git a/target/arm/cpu-param.h b/target/arm/cpu-param.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu-param.h +++ b/target/arm/cpu-param.h @@ -XXX,XX +XXX,XX @@ # define TARGET_PAGE_BITS_MIN 10 #endif -#define NB_MMU_MODES 12 +#define NB_MMU_MODES 11 #endif 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 @@ bool write_cpustate_to_list(ARMCPU *cpu, bool kvm_sync); * handling via the TLB. The only way to do a stage 1 translation without * the immediate stage 2 translation is via the ATS or AT system insns, * which can be slow-pathed and always do a page table walk. + * The only use of stage 2 translations is either as part of an s1+2 + * lookup or when loading the descriptors during a stage 1 page table walk, + * and in both those cases we don't use the TLB. * 4. we can also safely fold together the "32 bit EL3" and "64 bit EL3" * translation regimes, because they map reasonably well to each other * and they can't both be active at the same time. @@ -XXX,XX +XXX,XX @@ bool write_cpustate_to_list(ARMCPU *cpu, bool kvm_sync); * NS EL1 EL1&0 stage 1+2 (aka NS PL1) * NS EL1 EL1&0 stage 1+2 +PAN * NS EL0 EL2&0 + * NS EL2 EL2&0 * NS EL2 EL2&0 +PAN * NS EL2 (aka NS PL2) * S EL0 EL1&0 (aka S PL0) * S EL1 EL1&0 (not used if EL3 is 32 bit) * S EL1 EL1&0 +PAN * S EL3 (aka S PL1) - * NS EL1&0 stage 2 * - * for a total of 12 different mmu_idx. + * for a total of 11 different mmu_idx. * * R profile CPUs have an MPU, but can use the same set of MMU indexes * as A profile. They only need to distinguish NS EL0 and NS EL1 (and @@ -XXX,XX +XXX,XX @@ bool write_cpustate_to_list(ARMCPU *cpu, bool kvm_sync); * are not quite the same -- different CPU types (most notably M profile * vs A/R profile) would like to use MMU indexes with different semantics, * but since we don't ever need to use all of those in a single CPU we - * can avoid setting NB_MMU_MODES to more than 8. The lower bits of + * can avoid having to set NB_MMU_MODES to "total number of A profile MMU + * modes + total number of M profile MMU modes". The lower bits of * ARMMMUIdx are the core TLB mmu index, and the higher bits are always * the same for any particular CPU. * Variables of type ARMMUIdx are always full values, and the core @@ -XXX,XX +XXX,XX @@ typedef enum ARMMMUIdx { ARMMMUIdx_SE10_1_PAN = 9 | ARM_MMU_IDX_A, ARMMMUIdx_SE3 = 10 | ARM_MMU_IDX_A, - ARMMMUIdx_Stage2 = 11 | ARM_MMU_IDX_A, - /* * These are not allocated TLBs and are used only for AT system * instructions or for the first stage of an S12 page table walk. @@ -XXX,XX +XXX,XX @@ typedef enum ARMMMUIdx { ARMMMUIdx_Stage1_E0 = 0 | ARM_MMU_IDX_NOTLB, ARMMMUIdx_Stage1_E1 = 1 | ARM_MMU_IDX_NOTLB, ARMMMUIdx_Stage1_E1_PAN = 2 | ARM_MMU_IDX_NOTLB, + /* + * Not allocated a TLB: used only for second stage of an S12 page + * table walk, or for descriptor loads during first stage of an S1 + * page table walk. Note that if we ever want to have a TLB for this + * then various TLB flush insns which currently are no-ops or flush + * only stage 1 MMU indexes will need to change to flush stage 2. + */ + ARMMMUIdx_Stage2 = 3 | ARM_MMU_IDX_NOTLB, /* * M-profile. @@ -XXX,XX +XXX,XX @@ typedef enum ARMMMUIdxBit { TO_CORE_BIT(SE10_1), TO_CORE_BIT(SE10_1_PAN), TO_CORE_BIT(SE3), - TO_CORE_BIT(Stage2), TO_CORE_BIT(MUser), TO_CORE_BIT(MPriv), 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 tlbiall_nsnh_write(CPUARMState *env, const ARMCPRegInfo *ri, tlb_flush_by_mmuidx(cs, ARMMMUIdxBit_E10_1 | ARMMMUIdxBit_E10_1_PAN | - ARMMMUIdxBit_E10_0 | - ARMMMUIdxBit_Stage2); + ARMMMUIdxBit_E10_0); } static void tlbiall_nsnh_is_write(CPUARMState *env, const ARMCPRegInfo *ri, @@ -XXX,XX +XXX,XX @@ static void tlbiall_nsnh_is_write(CPUARMState *env, const ARMCPRegInfo *ri, tlb_flush_by_mmuidx_all_cpus_synced(cs, ARMMMUIdxBit_E10_1 | ARMMMUIdxBit_E10_1_PAN | - ARMMMUIdxBit_E10_0 | - ARMMMUIdxBit_Stage2); + ARMMMUIdxBit_E10_0); } -static void tlbiipas2_write(CPUARMState *env, const ARMCPRegInfo *ri, - uint64_t value) -{ - /* Invalidate by IPA. This has to invalidate any structures that - * contain only stage 2 translation information, but does not need - * to apply to structures that contain combined stage 1 and stage 2 - * translation information. - * This must NOP if EL2 isn't implemented or SCR_EL3.NS is zero. - */ - CPUState *cs = env_cpu(env); - uint64_t pageaddr; - - if (!arm_feature(env, ARM_FEATURE_EL2) || !(env->cp15.scr_el3 & SCR_NS)) { - return; - } - - pageaddr = sextract64(value << 12, 0, 40); - - tlb_flush_page_by_mmuidx(cs, pageaddr, ARMMMUIdxBit_Stage2); -} - -static void tlbiipas2_is_write(CPUARMState *env, const ARMCPRegInfo *ri, - uint64_t value) -{ - CPUState *cs = env_cpu(env); - uint64_t pageaddr; - - if (!arm_feature(env, ARM_FEATURE_EL2) || !(env->cp15.scr_el3 & SCR_NS)) { - return; - } - - pageaddr = sextract64(value << 12, 0, 40); - - tlb_flush_page_by_mmuidx_all_cpus_synced(cs, pageaddr, - ARMMMUIdxBit_Stage2); -} static void tlbiall_hyp_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value) @@ -XXX,XX +XXX,XX @@ static void vttbr_write(CPUARMState *env, const ARMCPRegInfo *ri, tlb_flush_by_mmuidx(cs, ARMMMUIdxBit_E10_1 | ARMMMUIdxBit_E10_1_PAN | - ARMMMUIdxBit_E10_0 | - ARMMMUIdxBit_Stage2); + ARMMMUIdxBit_E10_0); raw_write(env, ri, value); } } @@ -XXX,XX +XXX,XX @@ static int alle1_tlbmask(CPUARMState *env) return ARMMMUIdxBit_SE10_1 | ARMMMUIdxBit_SE10_1_PAN | ARMMMUIdxBit_SE10_0; - } else if (arm_feature(env, ARM_FEATURE_EL2)) { - return ARMMMUIdxBit_E10_1 | - ARMMMUIdxBit_E10_1_PAN | - ARMMMUIdxBit_E10_0 | - ARMMMUIdxBit_Stage2; } else { return ARMMMUIdxBit_E10_1 | ARMMMUIdxBit_E10_1_PAN | @@ -XXX,XX +XXX,XX @@ static void tlbi_aa64_vae3is_write(CPUARMState *env, const ARMCPRegInfo *ri, ARMMMUIdxBit_SE3); } -static void tlbi_aa64_ipas2e1_write(CPUARMState *env, const ARMCPRegInfo *ri, - uint64_t value) -{ - /* Invalidate by IPA. This has to invalidate any structures that - * contain only stage 2 translation information, but does not need - * to apply to structures that contain combined stage 1 and stage 2 - * translation information. - * This must NOP if EL2 isn't implemented or SCR_EL3.NS is zero. - */ - ARMCPU *cpu = env_archcpu(env); - CPUState *cs = CPU(cpu); - uint64_t pageaddr; - - if (!arm_feature(env, ARM_FEATURE_EL2) || !(env->cp15.scr_el3 & SCR_NS)) { - return; - } - - pageaddr = sextract64(value << 12, 0, 48); - - tlb_flush_page_by_mmuidx(cs, pageaddr, ARMMMUIdxBit_Stage2); -} - -static void tlbi_aa64_ipas2e1is_write(CPUARMState *env, const ARMCPRegInfo *ri, - uint64_t value) -{ - CPUState *cs = env_cpu(env); - uint64_t pageaddr; - - if (!arm_feature(env, ARM_FEATURE_EL2) || !(env->cp15.scr_el3 & SCR_NS)) { - return; - } - - pageaddr = sextract64(value << 12, 0, 48); - - tlb_flush_page_by_mmuidx_all_cpus_synced(cs, pageaddr, - ARMMMUIdxBit_Stage2); -} - static CPAccessResult aa64_zva_access(CPUARMState *env, const ARMCPRegInfo *ri, bool isread) { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo v8_cp_reginfo[] = { .writefn = tlbi_aa64_vae1_write }, { .name = "TLBI_IPAS2E1IS", .state = ARM_CP_STATE_AA64, .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 0, .opc2 = 1, - .access = PL2_W, .type = ARM_CP_NO_RAW, - .writefn = tlbi_aa64_ipas2e1is_write }, + .access = PL2_W, .type = ARM_CP_NOP }, { .name = "TLBI_IPAS2LE1IS", .state = ARM_CP_STATE_AA64, .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 0, .opc2 = 5, - .access = PL2_W, .type = ARM_CP_NO_RAW, - .writefn = tlbi_aa64_ipas2e1is_write }, + .access = PL2_W, .type = ARM_CP_NOP }, { .name = "TLBI_ALLE1IS", .state = ARM_CP_STATE_AA64, .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 3, .opc2 = 4, .access = PL2_W, .type = ARM_CP_NO_RAW, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo v8_cp_reginfo[] = { .writefn = tlbi_aa64_alle1is_write }, { .name = "TLBI_IPAS2E1", .state = ARM_CP_STATE_AA64, .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 4, .opc2 = 1, - .access = PL2_W, .type = ARM_CP_NO_RAW, - .writefn = tlbi_aa64_ipas2e1_write }, + .access = PL2_W, .type = ARM_CP_NOP }, { .name = "TLBI_IPAS2LE1", .state = ARM_CP_STATE_AA64, .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 4, .opc2 = 5, - .access = PL2_W, .type = ARM_CP_NO_RAW, - .writefn = tlbi_aa64_ipas2e1_write }, + .access = PL2_W, .type = ARM_CP_NOP }, { .name = "TLBI_ALLE1", .state = ARM_CP_STATE_AA64, .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 7, .opc2 = 4, .access = PL2_W, .type = ARM_CP_NO_RAW, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo v8_cp_reginfo[] = { .writefn = tlbimva_hyp_is_write }, { .name = "TLBIIPAS2", .cp = 15, .opc1 = 4, .crn = 8, .crm = 4, .opc2 = 1, - .type = ARM_CP_NO_RAW, .access = PL2_W, - .writefn = tlbiipas2_write }, + .type = ARM_CP_NOP, .access = PL2_W }, { .name = "TLBIIPAS2IS", .cp = 15, .opc1 = 4, .crn = 8, .crm = 0, .opc2 = 1, - .type = ARM_CP_NO_RAW, .access = PL2_W, - .writefn = tlbiipas2_is_write }, + .type = ARM_CP_NOP, .access = PL2_W }, { .name = "TLBIIPAS2L", .cp = 15, .opc1 = 4, .crn = 8, .crm = 4, .opc2 = 5, - .type = ARM_CP_NO_RAW, .access = PL2_W, - .writefn = tlbiipas2_write }, + .type = ARM_CP_NOP, .access = PL2_W }, { .name = "TLBIIPAS2LIS", .cp = 15, .opc1 = 4, .crn = 8, .crm = 0, .opc2 = 5, - .type = ARM_CP_NO_RAW, .access = PL2_W, - .writefn = tlbiipas2_is_write }, + .type = ARM_CP_NOP, .access = PL2_W }, /* 32 bit cache operations */ { .name = "ICIALLUIS", .cp = 15, .opc1 = 0, .crn = 7, .crm = 1, .opc2 = 0, .type = ARM_CP_NOP, .access = PL1_W, .accessfn = aa64_cacheop_pou_access }, -- 2.20.1 The access_type argument to get_phys_addr_lpae() is an MMUAccessType; use the enum constant MMU_DATA_LOAD rather than a literal 0 when we call it in S1_ptw_translate(). Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200330210400.11724-3-peter.maydell@linaro.org --- target/arm/helper.c | 5 +++-- 1 file changed, 3 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 @@ static hwaddr S1_ptw_translate(CPUARMState *env, ARMMMUIdx mmu_idx, pcacheattrs = &cacheattrs; } - ret = get_phys_addr_lpae(env, addr, 0, ARMMMUIdx_Stage2, &s2pa, - &txattrs, &s2prot, &s2size, fi, pcacheattrs); + ret = get_phys_addr_lpae(env, addr, MMU_DATA_LOAD, ARMMMUIdx_Stage2, + &s2pa, &txattrs, &s2prot, &s2size, fi, + pcacheattrs); if (ret) { assert(fi->type != ARMFault_None); fi->s2addr = addr; -- 2.20.1 For ARMv8.2-TTS2UXN, the stage 2 page table walk wants to know whether the stage 1 access is for EL0 or not, because whether exec permission is given can depend on whether this is an EL0 or EL1 access. Add a new argument to get_phys_addr_lpae() so the call sites can pass this information in. Since get_phys_addr_lpae() doesn't already have a doc comment, add one so we have a place to put the documentation of the semantics of the new s1_is_el0 argument. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200330210400.11724-4-peter.maydell@linaro.org --- target/arm/helper.c | 29 ++++++++++++++++++++++++++++- 1 file changed, 28 insertions(+), 1 deletion(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static bool get_phys_addr_lpae(CPUARMState *env, target_ulong address, MMUAccessType access_type, ARMMMUIdx mmu_idx, + bool s1_is_el0, hwaddr *phys_ptr, MemTxAttrs *txattrs, int *prot, target_ulong *page_size_ptr, ARMMMUFaultInfo *fi, ARMCacheAttrs *cacheattrs); @@ -XXX,XX +XXX,XX @@ static hwaddr S1_ptw_translate(CPUARMState *env, ARMMMUIdx mmu_idx, } ret = get_phys_addr_lpae(env, addr, MMU_DATA_LOAD, ARMMMUIdx_Stage2, + false, &s2pa, &txattrs, &s2prot, &s2size, fi, pcacheattrs); if (ret) { @@ -XXX,XX +XXX,XX @@ static ARMVAParameters aa32_va_parameters(CPUARMState *env, uint32_t va, }; } +/** + * get_phys_addr_lpae: perform one stage of page table walk, LPAE format + * + * Returns false if the translation was successful. Otherwise, phys_ptr, attrs, + * prot and page_size may not be filled in, and the populated fsr value provides + * information on why the translation aborted, in the format of a long-format + * DFSR/IFSR fault register, with the following caveats: + * * the WnR bit is never set (the caller must do this). + * + * @env: CPUARMState + * @address: virtual address to get physical address for + * @access_type: MMU_DATA_LOAD, MMU_DATA_STORE or MMU_INST_FETCH + * @mmu_idx: MMU index indicating required translation regime + * @s1_is_el0: if @mmu_idx is ARMMMUIdx_Stage2 (so this is a stage 2 page table + * walk), must be true if this is stage 2 of a stage 1+2 walk for an + * EL0 access). If @mmu_idx is anything else, @s1_is_el0 is ignored. + * @phys_ptr: set to the physical address corresponding to the virtual address + * @attrs: set to the memory transaction attributes to use + * @prot: set to the permissions for the page containing phys_ptr + * @page_size_ptr: set to the size of the page containing phys_ptr + * @fi: set to fault info if the translation fails + * @cacheattrs: (if non-NULL) set to the cacheability/shareability attributes + */ static bool get_phys_addr_lpae(CPUARMState *env, target_ulong address, MMUAccessType access_type, ARMMMUIdx mmu_idx, + bool s1_is_el0, hwaddr *phys_ptr, MemTxAttrs *txattrs, int *prot, target_ulong *page_size_ptr, ARMMMUFaultInfo *fi, ARMCacheAttrs *cacheattrs) @@ -XXX,XX +XXX,XX @@ bool get_phys_addr(CPUARMState *env, target_ulong address, /* S1 is done. Now do S2 translation. */ ret = get_phys_addr_lpae(env, ipa, access_type, ARMMMUIdx_Stage2, + mmu_idx == ARMMMUIdx_E10_0, phys_ptr, attrs, &s2_prot, page_size, fi, cacheattrs != NULL ? &cacheattrs2 : NULL); @@ -XXX,XX +XXX,XX @@ bool get_phys_addr(CPUARMState *env, target_ulong address, } if (regime_using_lpae_format(env, mmu_idx)) { - return get_phys_addr_lpae(env, address, access_type, mmu_idx, + return get_phys_addr_lpae(env, address, access_type, mmu_idx, false, phys_ptr, attrs, prot, page_size, fi, cacheattrs); } else if (regime_sctlr(env, mmu_idx) & SCTLR_XP) { -- 2.20.1 The ARMv8.2-TTS2UXN feature extends the XN field in stage 2 translation table descriptors from just bit [54] to bits [54:53], allowing stage 2 to control execution permissions separately for EL0 and EL1. Implement the new semantics of the XN field and enable the feature for our 'max' CPU. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200330210400.11724-5-peter.maydell@linaro.org --- target/arm/cpu.h | 15 +++++++++++++++ target/arm/cpu.c | 1 + target/arm/cpu64.c | 2 ++ target/arm/helper.c | 37 +++++++++++++++++++++++++++++++------ 4 files changed, 49 insertions(+), 6 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 isar_feature_aa32_ccidx(const ARMISARegisters *id) return FIELD_EX32(id->id_mmfr4, ID_MMFR4, CCIDX) != 0; } +static inline bool isar_feature_aa32_tts2uxn(const ARMISARegisters *id) +{ + return FIELD_EX32(id->id_mmfr4, ID_MMFR4, XNX) != 0; +} + /* * 64-bit feature tests via id registers. */ @@ -XXX,XX +XXX,XX @@ static inline bool isar_feature_aa64_ccidx(const ARMISARegisters *id) return FIELD_EX64(id->id_aa64mmfr2, ID_AA64MMFR2, CCIDX) != 0; } +static inline bool isar_feature_aa64_tts2uxn(const ARMISARegisters *id) +{ + return FIELD_EX64(id->id_aa64mmfr1, ID_AA64MMFR1, XNX) != 0; +} + /* * Feature tests for "does this exist in either 32-bit or 64-bit?" */ @@ -XXX,XX +XXX,XX @@ static inline bool isar_feature_any_ccidx(const ARMISARegisters *id) return isar_feature_aa64_ccidx(id) || isar_feature_aa32_ccidx(id); } +static inline bool isar_feature_any_tts2uxn(const ARMISARegisters *id) +{ + return isar_feature_aa64_tts2uxn(id) || isar_feature_aa32_tts2uxn(id); +} + /* * Forward to the above feature tests given an ARMCPU pointer. */ 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_max_initfn(Object *obj) t = FIELD_DP32(t, ID_MMFR4, HPDS, 1); /* AA32HPD */ t = FIELD_DP32(t, ID_MMFR4, AC2, 1); /* ACTLR2, HACTLR2 */ t = FIELD_DP32(t, ID_MMFR4, CNP, 1); /* TTCNP */ + t = FIELD_DP32(t, ID_MMFR4, XNX, 1); /* TTS2UXN */ cpu->isar.id_mmfr4 = t; } #endif 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_AA64MMFR1, VH, 1); t = FIELD_DP64(t, ID_AA64MMFR1, PAN, 2); /* ATS1E1 */ t = FIELD_DP64(t, ID_AA64MMFR1, VMIDBITS, 2); /* VMID16 */ + t = FIELD_DP64(t, ID_AA64MMFR1, XNX, 1); /* TTS2UXN */ cpu->isar.id_aa64mmfr1 = t; t = cpu->isar.id_aa64mmfr2; @@ -XXX,XX +XXX,XX @@ static void aarch64_max_initfn(Object *obj) u = FIELD_DP32(u, ID_MMFR4, HPDS, 1); /* AA32HPD */ u = FIELD_DP32(u, ID_MMFR4, AC2, 1); /* ACTLR2, HACTLR2 */ u = FIELD_DP32(u, ID_MMFR4, CNP, 1); /* TTCNP */ + u = FIELD_DP32(u, ID_MMFR4, XNX, 1); /* TTS2UXN */ cpu->isar.id_mmfr4 = u; u = cpu->isar.id_aa64dfr0; 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 @@ simple_ap_to_rw_prot(CPUARMState *env, ARMMMUIdx mmu_idx, int ap) * * @env: CPUARMState * @s2ap: The 2-bit stage2 access permissions (S2AP) - * @xn: XN (execute-never) bit + * @xn: XN (execute-never) bits + * @s1_is_el0: true if this is S2 of an S1+2 walk for EL0 */ -static int get_S2prot(CPUARMState *env, int s2ap, int xn) +static int get_S2prot(CPUARMState *env, int s2ap, int xn, bool s1_is_el0) { int prot = 0; @@ -XXX,XX +XXX,XX @@ static int get_S2prot(CPUARMState *env, int s2ap, int xn) if (s2ap & 2) { prot |= PAGE_WRITE; } - if (!xn) { - if (arm_el_is_aa64(env, 2) || prot & PAGE_READ) { + + if (cpu_isar_feature(any_tts2uxn, env_archcpu(env))) { + switch (xn) { + case 0: prot |= PAGE_EXEC; + break; + case 1: + if (s1_is_el0) { + prot |= PAGE_EXEC; + } + break; + case 2: + break; + case 3: + if (!s1_is_el0) { + prot |= PAGE_EXEC; + } + break; + default: + g_assert_not_reached(); + } + } else { + if (!extract32(xn, 1, 1)) { + if (arm_el_is_aa64(env, 2) || prot & PAGE_READ) { + prot |= PAGE_EXEC; + } } } return prot; @@ -XXX,XX +XXX,XX @@ static bool get_phys_addr_lpae(CPUARMState *env, target_ulong address, } ap = extract32(attrs, 4, 2); - xn = extract32(attrs, 12, 1); if (mmu_idx == ARMMMUIdx_Stage2) { ns = true; - *prot = get_S2prot(env, ap, xn); + xn = extract32(attrs, 11, 2); + *prot = get_S2prot(env, ap, xn, s1_is_el0); } else { ns = extract32(attrs, 3, 1); + xn = extract32(attrs, 12, 1); pxn = extract32(attrs, 11, 1); *prot = get_S1prot(env, mmu_idx, aarch64, ap, ns, xn, pxn); } -- 2.20.1 In aarch64_max_initfn() we update both 32-bit and 64-bit ID registers. The intended pattern is that for 64-bit ID registers we use FIELD_DP64 and the uint64_t 't' register, while 32-bit ID registers use FIELD_DP32 and the uint32_t 'u' register. For ID_AA64DFR0 we accidentally used 'u', meaning that the top 32 bits of this 64-bit ID register would end up always zero. Luckily at the moment that's what they should be anyway, so this bug has no visible effects. Use the right-sized variable. Fixes: 3bec78447a958d481991 Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Laurent Desnogues <laurent.desnogues@gmail.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200423110915.10527-1-peter.maydell@linaro.org --- target/arm/cpu64.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) 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) u = FIELD_DP32(u, ID_MMFR4, XNX, 1); /* TTS2UXN */ cpu->isar.id_mmfr4 = u; - u = cpu->isar.id_aa64dfr0; - u = FIELD_DP64(u, ID_AA64DFR0, PMUVER, 5); /* v8.4-PMU */ - cpu->isar.id_aa64dfr0 = u; + t = cpu->isar.id_aa64dfr0; + t = FIELD_DP64(t, ID_AA64DFR0, PMUVER, 5); /* v8.4-PMU */ + cpu->isar.id_aa64dfr0 = t; u = cpu->isar.id_dfr0; u = FIELD_DP32(u, ID_DFR0, PERFMON, 5); /* v8.4-PMU */ -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> MIDR_EL1 is a 64-bit system register with the top 32-bit being RES0. Represent it in QEMU's ARMCPU struct with a uint64_t, not a uint32_t. This fixes an error when compiling with -Werror=conversion because we were manipulating the register value using a local uint64_t variable: target/arm/cpu64.c: In function ‘aarch64_max_initfn’: target/arm/cpu64.c:628:21: error: conversion from ‘uint64_t’ {aka ‘long unsigned int’} to ‘uint32_t’ {aka ‘unsigned int’} may change value [-Werror=conversion] 628 | cpu->midr = t; | ^ and future-proofs us against a possible future architecture change using some of the top 32 bits. Suggested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Suggested-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Laurent Desnogues <laurent.desnogues@gmail.com> Message-id: 20200428172634.29707-1-f4bug@amsat.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 2 +- target/arm/cpu.c | 2 +- 2 files changed, 2 insertions(+), 2 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 @@ struct ARMCPU { uint64_t id_aa64dfr0; uint64_t id_aa64dfr1; } isar; - uint32_t midr; + uint64_t midr; uint32_t revidr; uint32_t reset_fpsid; uint32_t ctr; diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ static const ARMCPUInfo arm_cpus[] = { static Property arm_cpu_properties[] = { DEFINE_PROP_BOOL("start-powered-off", ARMCPU, start_powered_off, false), DEFINE_PROP_UINT32("psci-conduit", ARMCPU, psci_conduit, 0), - DEFINE_PROP_UINT32("midr", ARMCPU, midr, 0), + DEFINE_PROP_UINT64("midr", ARMCPU, midr, 0), DEFINE_PROP_UINT64("mp-affinity", ARMCPU, mp_affinity, ARM64_AFFINITY_INVALID), DEFINE_PROP_INT32("node-id", ARMCPU, node_id, CPU_UNSET_NUMA_NODE_ID), -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Remove inclusion of arm_gicv3_common.h, this already gets included via xlnx-versal.h. Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-2-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/xlnx-versal.c | 1 - 1 file changed, 1 deletion(-) diff --git a/hw/arm/xlnx-versal.c b/hw/arm/xlnx-versal.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal.c +++ b/hw/arm/xlnx-versal.c @@ -XXX,XX +XXX,XX @@ #include "hw/arm/boot.h" #include "kvm_arm.h" #include "hw/misc/unimp.h" -#include "hw/intc/arm_gicv3_common.h" #include "hw/arm/xlnx-versal.h" #include "hw/char/pl011.h" -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Move misplaced comment. Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-3-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/xlnx-versal.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/arm/xlnx-versal.c b/hw/arm/xlnx-versal.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal.c +++ b/hw/arm/xlnx-versal.c @@ -XXX,XX +XXX,XX @@ static void versal_create_apu_cpus(Versal *s) obj = object_new(XLNX_VERSAL_ACPU_TYPE); if (!obj) { - /* Secondary CPUs start in PSCI powered-down state */ error_report("Unable to create apu.cpu[%d] of type %s", i, XLNX_VERSAL_ACPU_TYPE); exit(EXIT_FAILURE); @@ -XXX,XX +XXX,XX @@ static void versal_create_apu_cpus(Versal *s) object_property_set_int(obj, s->cfg.psci_conduit, "psci-conduit", &error_abort); if (i) { + /* Secondary CPUs start in PSCI powered-down state */ object_property_set_bool(obj, true, "start-powered-off", &error_abort); } -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Fix typo xlnx-ve -> xlnx-versal. Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-4-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/xlnx-versal-virt.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/arm/xlnx-versal-virt.c b/hw/arm/xlnx-versal-virt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal-virt.c +++ b/hw/arm/xlnx-versal-virt.c @@ -XXX,XX +XXX,XX @@ static void versal_virt_init(MachineState *machine) psci_conduit = QEMU_PSCI_CONDUIT_SMC; } - sysbus_init_child_obj(OBJECT(machine), "xlnx-ve", &s->soc, + sysbus_init_child_obj(OBJECT(machine), "xlnx-versal", &s->soc, sizeof(s->soc), TYPE_XLNX_VERSAL); object_property_set_link(OBJECT(&s->soc), OBJECT(machine->ram), "ddr", &error_abort); -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Embed the UARTs into the SoC type. Suggested-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-5-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/xlnx-versal.h | 3 ++- hw/arm/xlnx-versal.c | 12 ++++++------ 2 files changed, 8 insertions(+), 7 deletions(-) diff --git a/include/hw/arm/xlnx-versal.h b/include/hw/arm/xlnx-versal.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/xlnx-versal.h +++ b/include/hw/arm/xlnx-versal.h @@ -XXX,XX +XXX,XX @@ #include "hw/sysbus.h" #include "hw/arm/boot.h" #include "hw/intc/arm_gicv3.h" +#include "hw/char/pl011.h" #define TYPE_XLNX_VERSAL "xlnx-versal" #define XLNX_VERSAL(obj) OBJECT_CHECK(Versal, (obj), TYPE_XLNX_VERSAL) @@ -XXX,XX +XXX,XX @@ typedef struct Versal { MemoryRegion mr_ocm; struct { - SysBusDevice *uart[XLNX_VERSAL_NR_UARTS]; + PL011State uart[XLNX_VERSAL_NR_UARTS]; SysBusDevice *gem[XLNX_VERSAL_NR_GEMS]; SysBusDevice *adma[XLNX_VERSAL_NR_ADMAS]; } iou; diff --git a/hw/arm/xlnx-versal.c b/hw/arm/xlnx-versal.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal.c +++ b/hw/arm/xlnx-versal.c @@ -XXX,XX +XXX,XX @@ #include "kvm_arm.h" #include "hw/misc/unimp.h" #include "hw/arm/xlnx-versal.h" -#include "hw/char/pl011.h" #define XLNX_VERSAL_ACPU_TYPE ARM_CPU_TYPE_NAME("cortex-a72") #define GEM_REVISION 0x40070106 @@ -XXX,XX +XXX,XX @@ static void versal_create_uarts(Versal *s, qemu_irq *pic) DeviceState *dev; MemoryRegion *mr; - dev = qdev_create(NULL, TYPE_PL011); - s->lpd.iou.uart[i] = SYS_BUS_DEVICE(dev); + sysbus_init_child_obj(OBJECT(s), name, + &s->lpd.iou.uart[i], sizeof(s->lpd.iou.uart[i]), + TYPE_PL011); + dev = DEVICE(&s->lpd.iou.uart[i]); qdev_prop_set_chr(dev, "chardev", serial_hd(i)); - object_property_add_child(OBJECT(s), name, OBJECT(dev), &error_fatal); qdev_init_nofail(dev); - mr = sysbus_mmio_get_region(s->lpd.iou.uart[i], 0); + mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0); memory_region_add_subregion(&s->mr_ps, addrs[i], mr); - sysbus_connect_irq(s->lpd.iou.uart[i], 0, pic[irqs[i]]); + sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, pic[irqs[i]]); g_free(name); } } -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Embed the GEMs into the SoC type. Suggested-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-6-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/xlnx-versal.h | 3 ++- hw/arm/xlnx-versal.c | 15 ++++++++------- 2 files changed, 10 insertions(+), 8 deletions(-) diff --git a/include/hw/arm/xlnx-versal.h b/include/hw/arm/xlnx-versal.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/xlnx-versal.h +++ b/include/hw/arm/xlnx-versal.h @@ -XXX,XX +XXX,XX @@ #include "hw/arm/boot.h" #include "hw/intc/arm_gicv3.h" #include "hw/char/pl011.h" +#include "hw/net/cadence_gem.h" #define TYPE_XLNX_VERSAL "xlnx-versal" #define XLNX_VERSAL(obj) OBJECT_CHECK(Versal, (obj), TYPE_XLNX_VERSAL) @@ -XXX,XX +XXX,XX @@ typedef struct Versal { struct { PL011State uart[XLNX_VERSAL_NR_UARTS]; - SysBusDevice *gem[XLNX_VERSAL_NR_GEMS]; + CadenceGEMState gem[XLNX_VERSAL_NR_GEMS]; SysBusDevice *adma[XLNX_VERSAL_NR_ADMAS]; } iou; } lpd; diff --git a/hw/arm/xlnx-versal.c b/hw/arm/xlnx-versal.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal.c +++ b/hw/arm/xlnx-versal.c @@ -XXX,XX +XXX,XX @@ static void versal_create_gems(Versal *s, qemu_irq *pic) DeviceState *dev; MemoryRegion *mr; - dev = qdev_create(NULL, "cadence_gem"); - s->lpd.iou.gem[i] = SYS_BUS_DEVICE(dev); - object_property_add_child(OBJECT(s), name, OBJECT(dev), &error_fatal); + sysbus_init_child_obj(OBJECT(s), name, + &s->lpd.iou.gem[i], sizeof(s->lpd.iou.gem[i]), + TYPE_CADENCE_GEM); + dev = DEVICE(&s->lpd.iou.gem[i]); if (nd->used) { qemu_check_nic_model(nd, "cadence_gem"); qdev_set_nic_properties(dev, nd); } - object_property_set_int(OBJECT(s->lpd.iou.gem[i]), + object_property_set_int(OBJECT(dev), 2, "num-priority-queues", &error_abort); - object_property_set_link(OBJECT(s->lpd.iou.gem[i]), + object_property_set_link(OBJECT(dev), OBJECT(&s->mr_ps), "dma", &error_abort); qdev_init_nofail(dev); - mr = sysbus_mmio_get_region(s->lpd.iou.gem[i], 0); + mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0); memory_region_add_subregion(&s->mr_ps, addrs[i], mr); - sysbus_connect_irq(s->lpd.iou.gem[i], 0, pic[irqs[i]]); + sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, pic[irqs[i]]); g_free(name); } } -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Embed the ADMAs into the SoC type. Suggested-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-7-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/xlnx-versal.h | 3 ++- hw/arm/xlnx-versal.c | 14 +++++++------- 2 files changed, 9 insertions(+), 8 deletions(-) diff --git a/include/hw/arm/xlnx-versal.h b/include/hw/arm/xlnx-versal.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/xlnx-versal.h +++ b/include/hw/arm/xlnx-versal.h @@ -XXX,XX +XXX,XX @@ #include "hw/arm/boot.h" #include "hw/intc/arm_gicv3.h" #include "hw/char/pl011.h" +#include "hw/dma/xlnx-zdma.h" #include "hw/net/cadence_gem.h" #define TYPE_XLNX_VERSAL "xlnx-versal" @@ -XXX,XX +XXX,XX @@ typedef struct Versal { struct { PL011State uart[XLNX_VERSAL_NR_UARTS]; CadenceGEMState gem[XLNX_VERSAL_NR_GEMS]; - SysBusDevice *adma[XLNX_VERSAL_NR_ADMAS]; + XlnxZDMA adma[XLNX_VERSAL_NR_ADMAS]; } iou; } lpd; diff --git a/hw/arm/xlnx-versal.c b/hw/arm/xlnx-versal.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal.c +++ b/hw/arm/xlnx-versal.c @@ -XXX,XX +XXX,XX @@ static void versal_create_admas(Versal *s, qemu_irq *pic) DeviceState *dev; MemoryRegion *mr; - dev = qdev_create(NULL, "xlnx.zdma"); - s->lpd.iou.adma[i] = SYS_BUS_DEVICE(dev); - object_property_set_int(OBJECT(s->lpd.iou.adma[i]), 128, "bus-width", - &error_abort); - object_property_add_child(OBJECT(s), name, OBJECT(dev), &error_fatal); + sysbus_init_child_obj(OBJECT(s), name, + &s->lpd.iou.adma[i], sizeof(s->lpd.iou.adma[i]), + TYPE_XLNX_ZDMA); + dev = DEVICE(&s->lpd.iou.adma[i]); + object_property_set_int(OBJECT(dev), 128, "bus-width", &error_abort); qdev_init_nofail(dev); - mr = sysbus_mmio_get_region(s->lpd.iou.adma[i], 0); + mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0); memory_region_add_subregion(&s->mr_ps, MM_ADMA_CH0 + i * MM_ADMA_CH0_SIZE, mr); - sysbus_connect_irq(s->lpd.iou.adma[i], 0, pic[VERSAL_ADMA_IRQ_0 + i]); + sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, pic[VERSAL_ADMA_IRQ_0 + i]); g_free(name); } } -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Embed the APUs into the SoC type. Suggested-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-8-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/xlnx-versal.h | 2 +- hw/arm/xlnx-versal-virt.c | 4 ++-- hw/arm/xlnx-versal.c | 19 +++++-------------- 3 files changed, 8 insertions(+), 17 deletions(-) diff --git a/include/hw/arm/xlnx-versal.h b/include/hw/arm/xlnx-versal.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/xlnx-versal.h +++ b/include/hw/arm/xlnx-versal.h @@ -XXX,XX +XXX,XX @@ typedef struct Versal { struct { struct { MemoryRegion mr; - ARMCPU *cpu[XLNX_VERSAL_NR_ACPUS]; + ARMCPU cpu[XLNX_VERSAL_NR_ACPUS]; GICv3State gic; } apu; } fpd; diff --git a/hw/arm/xlnx-versal-virt.c b/hw/arm/xlnx-versal-virt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal-virt.c +++ b/hw/arm/xlnx-versal-virt.c @@ -XXX,XX +XXX,XX @@ static void versal_virt_init(MachineState *machine) s->binfo.get_dtb = versal_virt_get_dtb; s->binfo.modify_dtb = versal_virt_modify_dtb; if (machine->kernel_filename) { - arm_load_kernel(s->soc.fpd.apu.cpu[0], machine, &s->binfo); + arm_load_kernel(&s->soc.fpd.apu.cpu[0], machine, &s->binfo); } else { - AddressSpace *as = arm_boot_address_space(s->soc.fpd.apu.cpu[0], + AddressSpace *as = arm_boot_address_space(&s->soc.fpd.apu.cpu[0], &s->binfo); /* Some boot-loaders (e.g u-boot) don't like blobs at address 0 (NULL). * Offset things by 4K. */ diff --git a/hw/arm/xlnx-versal.c b/hw/arm/xlnx-versal.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal.c +++ b/hw/arm/xlnx-versal.c @@ -XXX,XX +XXX,XX @@ static void versal_create_apu_cpus(Versal *s) for (i = 0; i < ARRAY_SIZE(s->fpd.apu.cpu); i++) { Object *obj; - char *name; - - obj = object_new(XLNX_VERSAL_ACPU_TYPE); - if (!obj) { - error_report("Unable to create apu.cpu[%d] of type %s", - i, XLNX_VERSAL_ACPU_TYPE); - exit(EXIT_FAILURE); - } - - name = g_strdup_printf("apu-cpu[%d]", i); - object_property_add_child(OBJECT(s), name, obj, &error_fatal); - g_free(name); + object_initialize_child(OBJECT(s), "apu-cpu[*]", + &s->fpd.apu.cpu[i], sizeof(s->fpd.apu.cpu[i]), + XLNX_VERSAL_ACPU_TYPE, &error_abort, NULL); + obj = OBJECT(&s->fpd.apu.cpu[i]); object_property_set_int(obj, s->cfg.psci_conduit, "psci-conduit", &error_abort); if (i) { @@ -XXX,XX +XXX,XX @@ static void versal_create_apu_cpus(Versal *s) object_property_set_link(obj, OBJECT(&s->fpd.apu.mr), "memory", &error_abort); object_property_set_bool(obj, true, "realized", &error_fatal); - s->fpd.apu.cpu[i] = ARM_CPU(obj); } } @@ -XXX,XX +XXX,XX @@ static void versal_create_apu_gic(Versal *s, qemu_irq *pic) } for (i = 0; i < nr_apu_cpus; i++) { - DeviceState *cpudev = DEVICE(s->fpd.apu.cpu[i]); + DeviceState *cpudev = DEVICE(&s->fpd.apu.cpu[i]); int ppibase = XLNX_VERSAL_NR_IRQS + i * GIC_INTERNAL + GIC_NR_SGIS; qemu_irq maint_irq; int ti; -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Add support for SD. Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-9-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/xlnx-versal.h | 12 ++++++++++++ hw/arm/xlnx-versal.c | 31 +++++++++++++++++++++++++++++++ 2 files changed, 43 insertions(+) diff --git a/include/hw/arm/xlnx-versal.h b/include/hw/arm/xlnx-versal.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/xlnx-versal.h +++ b/include/hw/arm/xlnx-versal.h @@ -XXX,XX +XXX,XX @@ #include "hw/sysbus.h" #include "hw/arm/boot.h" +#include "hw/sd/sdhci.h" #include "hw/intc/arm_gicv3.h" #include "hw/char/pl011.h" #include "hw/dma/xlnx-zdma.h" @@ -XXX,XX +XXX,XX @@ #define XLNX_VERSAL_NR_UARTS 2 #define XLNX_VERSAL_NR_GEMS 2 #define XLNX_VERSAL_NR_ADMAS 8 +#define XLNX_VERSAL_NR_SDS 2 #define XLNX_VERSAL_NR_IRQS 192 typedef struct Versal { @@ -XXX,XX +XXX,XX @@ typedef struct Versal { } iou; } lpd; + /* The Platform Management Controller subsystem. */ + struct { + struct { + SDHCIState sd[XLNX_VERSAL_NR_SDS]; + } iou; + } pmc; + struct { MemoryRegion *mr_ddr; uint32_t psci_conduit; @@ -XXX,XX +XXX,XX @@ typedef struct Versal { #define VERSAL_GEM1_IRQ_0 58 #define VERSAL_GEM1_WAKE_IRQ_0 59 #define VERSAL_ADMA_IRQ_0 60 +#define VERSAL_SD0_IRQ_0 126 /* Architecturally reserved IRQs suitable for virtualization. */ #define VERSAL_RSVD_IRQ_FIRST 111 @@ -XXX,XX +XXX,XX @@ typedef struct Versal { #define MM_FPD_CRF 0xfd1a0000U #define MM_FPD_CRF_SIZE 0x140000 +#define MM_PMC_SD0 0xf1040000U +#define MM_PMC_SD0_SIZE 0x10000 #define MM_PMC_CRP 0xf1260000U #define MM_PMC_CRP_SIZE 0x10000 #endif diff --git a/hw/arm/xlnx-versal.c b/hw/arm/xlnx-versal.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal.c +++ b/hw/arm/xlnx-versal.c @@ -XXX,XX +XXX,XX @@ static void versal_create_admas(Versal *s, qemu_irq *pic) } } +#define SDHCI_CAPABILITIES 0x280737ec6481 /* Same as on ZynqMP. */ +static void versal_create_sds(Versal *s, qemu_irq *pic) +{ + int i; + + for (i = 0; i < ARRAY_SIZE(s->pmc.iou.sd); i++) { + DeviceState *dev; + MemoryRegion *mr; + + sysbus_init_child_obj(OBJECT(s), "sd[*]", + &s->pmc.iou.sd[i], sizeof(s->pmc.iou.sd[i]), + TYPE_SYSBUS_SDHCI); + dev = DEVICE(&s->pmc.iou.sd[i]); + + object_property_set_uint(OBJECT(dev), + 3, "sd-spec-version", &error_fatal); + object_property_set_uint(OBJECT(dev), SDHCI_CAPABILITIES, "capareg", + &error_fatal); + object_property_set_uint(OBJECT(dev), UHS_I, "uhs", &error_fatal); + qdev_init_nofail(dev); + + mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0); + memory_region_add_subregion(&s->mr_ps, + MM_PMC_SD0 + i * MM_PMC_SD0_SIZE, mr); + + sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, + pic[VERSAL_SD0_IRQ_0 + i * 2]); + } +} + /* This takes the board allocated linear DDR memory and creates aliases * for each split DDR range/aperture on the Versal address map. */ @@ -XXX,XX +XXX,XX @@ static void versal_realize(DeviceState *dev, Error **errp) versal_create_uarts(s, pic); versal_create_gems(s, pic); versal_create_admas(s, pic); + versal_create_sds(s, pic); versal_map_ddr(s); versal_unimp(s); -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> hw/arm: versal: Add support for the RTC. Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-10-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/xlnx-versal.h | 8 ++++++++ hw/arm/xlnx-versal.c | 21 +++++++++++++++++++++ 2 files changed, 29 insertions(+) diff --git a/include/hw/arm/xlnx-versal.h b/include/hw/arm/xlnx-versal.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/xlnx-versal.h +++ b/include/hw/arm/xlnx-versal.h @@ -XXX,XX +XXX,XX @@ #include "hw/char/pl011.h" #include "hw/dma/xlnx-zdma.h" #include "hw/net/cadence_gem.h" +#include "hw/rtc/xlnx-zynqmp-rtc.h" #define TYPE_XLNX_VERSAL "xlnx-versal" #define XLNX_VERSAL(obj) OBJECT_CHECK(Versal, (obj), TYPE_XLNX_VERSAL) @@ -XXX,XX +XXX,XX @@ typedef struct Versal { struct { SDHCIState sd[XLNX_VERSAL_NR_SDS]; } iou; + + XlnxZynqMPRTC rtc; } pmc; struct { @@ -XXX,XX +XXX,XX @@ typedef struct Versal { #define VERSAL_GEM1_IRQ_0 58 #define VERSAL_GEM1_WAKE_IRQ_0 59 #define VERSAL_ADMA_IRQ_0 60 +#define VERSAL_RTC_APB_ERR_IRQ 121 #define VERSAL_SD0_IRQ_0 126 +#define VERSAL_RTC_ALARM_IRQ 142 +#define VERSAL_RTC_SECONDS_IRQ 143 /* Architecturally reserved IRQs suitable for virtualization. */ #define VERSAL_RSVD_IRQ_FIRST 111 @@ -XXX,XX +XXX,XX @@ typedef struct Versal { #define MM_PMC_SD0_SIZE 0x10000 #define MM_PMC_CRP 0xf1260000U #define MM_PMC_CRP_SIZE 0x10000 +#define MM_PMC_RTC 0xf12a0000 +#define MM_PMC_RTC_SIZE 0x10000 #endif diff --git a/hw/arm/xlnx-versal.c b/hw/arm/xlnx-versal.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal.c +++ b/hw/arm/xlnx-versal.c @@ -XXX,XX +XXX,XX @@ static void versal_create_sds(Versal *s, qemu_irq *pic) } } +static void versal_create_rtc(Versal *s, qemu_irq *pic) +{ + SysBusDevice *sbd; + MemoryRegion *mr; + + sysbus_init_child_obj(OBJECT(s), "rtc", &s->pmc.rtc, sizeof(s->pmc.rtc), + TYPE_XLNX_ZYNQMP_RTC); + sbd = SYS_BUS_DEVICE(&s->pmc.rtc); + qdev_init_nofail(DEVICE(sbd)); + + mr = sysbus_mmio_get_region(sbd, 0); + memory_region_add_subregion(&s->mr_ps, MM_PMC_RTC, mr); + + /* + * TODO: Connect the ALARM and SECONDS interrupts once our RTC model + * supports them. + */ + sysbus_connect_irq(sbd, 1, pic[VERSAL_RTC_APB_ERR_IRQ]); +} + /* This takes the board allocated linear DDR memory and creates aliases * for each split DDR range/aperture on the Versal address map. */ @@ -XXX,XX +XXX,XX @@ static void versal_realize(DeviceState *dev, Error **errp) versal_create_gems(s, pic); versal_create_admas(s, pic); versal_create_sds(s, pic); + versal_create_rtc(s, pic); versal_map_ddr(s); versal_unimp(s); -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Add support for SD. Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-11-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/xlnx-versal-virt.c | 46 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 46 insertions(+) diff --git a/hw/arm/xlnx-versal-virt.c b/hw/arm/xlnx-versal-virt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal-virt.c +++ b/hw/arm/xlnx-versal-virt.c @@ -XXX,XX +XXX,XX @@ #include "hw/arm/sysbus-fdt.h" #include "hw/arm/fdt.h" #include "cpu.h" +#include "hw/qdev-properties.h" #include "hw/arm/xlnx-versal.h" #define TYPE_XLNX_VERSAL_VIRT_MACHINE MACHINE_TYPE_NAME("xlnx-versal-virt") @@ -XXX,XX +XXX,XX @@ static void fdt_add_zdma_nodes(VersalVirt *s) } } +static void fdt_add_sd_nodes(VersalVirt *s) +{ + const char clocknames[] = "clk_xin\0clk_ahb"; + const char compat[] = "arasan,sdhci-8.9a"; + int i; + + for (i = ARRAY_SIZE(s->soc.pmc.iou.sd) - 1; i >= 0; i--) { + uint64_t addr = MM_PMC_SD0 + MM_PMC_SD0_SIZE * i; + char *name = g_strdup_printf("/sdhci@%" PRIx64, addr); + + qemu_fdt_add_subnode(s->fdt, name); + + qemu_fdt_setprop_cells(s->fdt, name, "clocks", + s->phandle.clk_25Mhz, s->phandle.clk_25Mhz); + qemu_fdt_setprop(s->fdt, name, "clock-names", + clocknames, sizeof(clocknames)); + qemu_fdt_setprop_cells(s->fdt, name, "interrupts", + GIC_FDT_IRQ_TYPE_SPI, VERSAL_SD0_IRQ_0 + i * 2, + GIC_FDT_IRQ_FLAGS_LEVEL_HI); + qemu_fdt_setprop_sized_cells(s->fdt, name, "reg", + 2, addr, 2, MM_PMC_SD0_SIZE); + qemu_fdt_setprop(s->fdt, name, "compatible", compat, sizeof(compat)); + g_free(name); + } +} + static void fdt_nop_memory_nodes(void *fdt, Error **errp) { Error *err = NULL; @@ -XXX,XX +XXX,XX @@ static void create_virtio_regions(VersalVirt *s) } } +static void sd_plugin_card(SDHCIState *sd, DriveInfo *di) +{ + BlockBackend *blk = di ? blk_by_legacy_dinfo(di) : NULL; + DeviceState *card; + + card = qdev_create(qdev_get_child_bus(DEVICE(sd), "sd-bus"), TYPE_SD_CARD); + object_property_add_child(OBJECT(sd), "card[*]", OBJECT(card), + &error_fatal); + qdev_prop_set_drive(card, "drive", blk, &error_fatal); + object_property_set_bool(OBJECT(card), true, "realized", &error_fatal); +} + static void versal_virt_init(MachineState *machine) { VersalVirt *s = XLNX_VERSAL_VIRT_MACHINE(machine); int psci_conduit = QEMU_PSCI_CONDUIT_DISABLED; + int i; /* * If the user provides an Operating System to be loaded, we expect them @@ -XXX,XX +XXX,XX @@ static void versal_virt_init(MachineState *machine) fdt_add_gic_nodes(s); fdt_add_timer_nodes(s); fdt_add_zdma_nodes(s); + fdt_add_sd_nodes(s); fdt_add_cpu_nodes(s, psci_conduit); fdt_add_clk_node(s, "/clk125", 125000000, s->phandle.clk_125Mhz); fdt_add_clk_node(s, "/clk25", 25000000, s->phandle.clk_25Mhz); @@ -XXX,XX +XXX,XX @@ static void versal_virt_init(MachineState *machine) memory_region_add_subregion_overlap(get_system_memory(), 0, &s->soc.fpd.apu.mr, 0); + /* Plugin SD cards. */ + for (i = 0; i < ARRAY_SIZE(s->soc.pmc.iou.sd); i++) { + sd_plugin_card(&s->soc.pmc.iou.sd[i], drive_get_next(IF_SD)); + } + s->binfo.ram_size = machine->ram_size; s->binfo.loader_start = 0x0; s->binfo.get_dtb = versal_virt_get_dtb; -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Add support for the RTC. Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-12-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/xlnx-versal-virt.c | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+) diff --git a/hw/arm/xlnx-versal-virt.c b/hw/arm/xlnx-versal-virt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal-virt.c +++ b/hw/arm/xlnx-versal-virt.c @@ -XXX,XX +XXX,XX @@ static void fdt_add_sd_nodes(VersalVirt *s) } } +static void fdt_add_rtc_node(VersalVirt *s) +{ + const char compat[] = "xlnx,zynqmp-rtc"; + const char interrupt_names[] = "alarm\0sec"; + char *name = g_strdup_printf("/rtc@%x", MM_PMC_RTC); + + qemu_fdt_add_subnode(s->fdt, name); + + qemu_fdt_setprop_cells(s->fdt, name, "interrupts", + GIC_FDT_IRQ_TYPE_SPI, VERSAL_RTC_ALARM_IRQ, + GIC_FDT_IRQ_FLAGS_LEVEL_HI, + GIC_FDT_IRQ_TYPE_SPI, VERSAL_RTC_SECONDS_IRQ, + GIC_FDT_IRQ_FLAGS_LEVEL_HI); + qemu_fdt_setprop(s->fdt, name, "interrupt-names", + interrupt_names, sizeof(interrupt_names)); + qemu_fdt_setprop_sized_cells(s->fdt, name, "reg", + 2, MM_PMC_RTC, 2, MM_PMC_RTC_SIZE); + qemu_fdt_setprop(s->fdt, name, "compatible", compat, sizeof(compat)); + g_free(name); +} + static void fdt_nop_memory_nodes(void *fdt, Error **errp) { Error *err = NULL; @@ -XXX,XX +XXX,XX @@ static void versal_virt_init(MachineState *machine) fdt_add_timer_nodes(s); fdt_add_zdma_nodes(s); fdt_add_sd_nodes(s); + fdt_add_rtc_node(s); fdt_add_cpu_nodes(s, psci_conduit); fdt_add_clk_node(s, "/clk125", 125000000, s->phandle.clk_125Mhz); fdt_add_clk_node(s, "/clk25", 25000000, s->phandle.clk_25Mhz); -- 2.20.1 Somewhere along theline we accidentally added a duplicate "using D16-D31 when they don't exist" check to do_vfm_dp() (probably an artifact of a patchseries rebase). Remove it. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200430181003.21682-2-peter.maydell@linaro.org --- target/arm/translate-vfp.inc.c | 6 ------ 1 file changed, 6 deletions(-) diff --git a/target/arm/translate-vfp.inc.c b/target/arm/translate-vfp.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-vfp.inc.c +++ b/target/arm/translate-vfp.inc.c @@ -XXX,XX +XXX,XX @@ static bool do_vfm_dp(DisasContext *s, arg_VFMA_dp *a, bool neg_n, bool neg_d) return false; } - /* UNDEF accesses to D16-D31 if they don't exist. */ - if (!dc_isar_feature(aa32_simd_r32, s) && - ((a->vd | a->vn | a->vm) & 0x10)) { - return false; - } - if (!vfp_access_check(s)) { return true; } -- 2.20.1 We were accidentally permitting decode of Thumb Neon insns even if the CPU didn't have the FEATURE_NEON bit set, because the feature check was being done before the call to disas_neon_data_insn() and disas_neon_ls_insn() in the Arm decoder but was omitted from the Thumb decoder. Push the feature bit check down into the called functions so it is done for both Arm and Thumb encodings. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200430181003.21682-3-peter.maydell@linaro.org --- target/arm/translate.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_ls_insn(DisasContext *s, uint32_t insn) TCGv_i32 tmp2; TCGv_i64 tmp64; + if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { + return 1; + } + /* FIXME: this access check should not take precedence over UNDEF * for invalid encodings; we will generate incorrect syndrome information * for attempts to execute invalid vfp/neon encodings with FP disabled. @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) TCGv_ptr ptr1, ptr2, ptr3; TCGv_i64 tmp64; + if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { + return 1; + } + /* FIXME: this access check should not take precedence over UNDEF * for invalid encodings; we will generate incorrect syndrome information * for attempts to execute invalid vfp/neon encodings with FP disabled. @@ -XXX,XX +XXX,XX @@ static void disas_arm_insn(DisasContext *s, unsigned int insn) if (((insn >> 25) & 7) == 1) { /* NEON Data processing. */ - if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { - goto illegal_op; - } - if (disas_neon_data_insn(s, insn)) { goto illegal_op; } @@ -XXX,XX +XXX,XX @@ static void disas_arm_insn(DisasContext *s, unsigned int insn) } if ((insn & 0x0f100000) == 0x04000000) { /* NEON load/store. */ - if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { - goto illegal_op; - } - if (disas_neon_ls_insn(s, insn)) { goto illegal_op; } -- 2.20.1 Add the infrastructure for building and invoking a decodetree decoder for the AArch32 Neon encodings. At the moment the new decoder covers nothing, so we always fall back to the existing hand-written decode. We follow the same pattern we did for the VFP decodetree conversion (commit 78e138bc1f672c145ef6ace74617d and following): code that deals with Neon will be moving gradually out to translate-neon.vfp.inc, which we #include into translate.c. In order to share the decode files between A32 and T32, we split Neon into 3 parts: * data-processing * load-store * 'shared' encodings The first two groups of instructions have similar but not identical A32 and T32 encodings, so we need to manually transform the T32 encoding into the A32 one before calling the decoder; the third group covers the Neon instructions which are identical in A32 and T32. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-4-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 29 ++++++++++++++++++++++++++ target/arm/neon-ls.decode | 29 ++++++++++++++++++++++++++ target/arm/neon-shared.decode | 27 +++++++++++++++++++++++++ target/arm/translate-neon.inc.c | 32 +++++++++++++++++++++++++++++ target/arm/translate.c | 36 +++++++++++++++++++++++++++++++-- target/arm/Makefile.objs | 18 +++++++++++++++++ 6 files changed, 169 insertions(+), 2 deletions(-) create mode 100644 target/arm/neon-dp.decode create mode 100644 target/arm/neon-ls.decode create mode 100644 target/arm/neon-shared.decode create mode 100644 target/arm/translate-neon.inc.c diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ +# AArch32 Neon data-processing instruction descriptions +# +# Copyright (c) 2020 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 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 +# + +# Encodings for Neon data processing instructions where the T32 encoding +# is a simple transformation of the A32 encoding. +# More specifically, this file covers instructions where the A32 encoding is +# 0b1111_001p_qqqq_qqqq_qqqq_qqqq_qqqq_qqqq +# and the T32 encoding is +# 0b111p_1111_qqqq_qqqq_qqqq_qqqq_qqqq_qqqq +# This file works on the A32 encoding only; calling code for T32 has to +# transform the insn into the A32 version first. diff --git a/target/arm/neon-ls.decode b/target/arm/neon-ls.decode new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/neon-ls.decode @@ -XXX,XX +XXX,XX @@ +# AArch32 Neon load/store instruction descriptions +# +# Copyright (c) 2020 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 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 +# + +# Encodings for Neon load/store instructions where the T32 encoding +# is a simple transformation of the A32 encoding. +# More specifically, this file covers instructions where the A32 encoding is +# 0b1111_0100_xxx0_xxxx_xxxx_xxxx_xxxx_xxxx +# and the T32 encoding is +# 0b1111_1001_xxx0_xxxx_xxxx_xxxx_xxxx_xxxx +# This file works on the A32 encoding only; calling code for T32 has to +# transform the insn into the A32 version first. diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ +# AArch32 Neon instruction descriptions +# +# Copyright (c) 2020 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 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 +# + +# Encodings for Neon instructions whose encoding is the same for +# both A32 and T32. + +# More specifically, this covers: +# 2reg scalar ext: 0b1111_1110_xxxx_xxxx_xxxx_1x0x_xxxx_xxxx +# 3same ext: 0b1111_110x_xxxx_xxxx_xxxx_1x0x_xxxx_xxxx diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ +/* + * ARM translation: AArch32 Neon instructions + * + * Copyright (c) 2003 Fabrice Bellard + * Copyright (c) 2005-2007 CodeSourcery + * Copyright (c) 2007 OpenedHand, Ltd. + * Copyright (c) 2020 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 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 intended to be included from translate.c; it uses + * some macros and definitions provided by that file. + * It might be possible to convert it to a standalone .c file eventually. + */ + +/* Include the generated Neon decoder */ +#include "decode-neon-dp.inc.c" +#include "decode-neon-ls.inc.c" +#include "decode-neon-shared.inc.c" 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 TCGv_ptr vfp_reg_ptr(bool dp, int reg) #define ARM_CP_RW_BIT (1 << 20) -/* Include the VFP decoder */ +/* Include the VFP and Neon decoders */ #include "translate-vfp.inc.c" +#include "translate-neon.inc.c" static inline void iwmmxt_load_reg(TCGv_i64 var, int reg) { @@ -XXX,XX +XXX,XX @@ static void disas_arm_insn(DisasContext *s, unsigned int insn) /* Unconditional instructions. */ /* TODO: Perhaps merge these into one decodetree output file. */ if (disas_a32_uncond(s, insn) || - disas_vfp_uncond(s, insn)) { + disas_vfp_uncond(s, insn) || + disas_neon_dp(s, insn) || + disas_neon_ls(s, insn) || + disas_neon_shared(s, insn)) { return; } /* fall back to legacy decoder */ @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) ARCH(6T2); } + if ((insn & 0xef000000) == 0xef000000) { + /* + * T32 encodings 0b111p_1111_qqqq_qqqq_qqqq_qqqq_qqqq_qqqq + * transform into + * A32 encodings 0b1111_001p_qqqq_qqqq_qqqq_qqqq_qqqq_qqqq + */ + uint32_t a32_insn = (insn & 0xe2ffffff) | + ((insn & (1 << 28)) >> 4) | (1 << 28); + + if (disas_neon_dp(s, a32_insn)) { + return; + } + } + + if ((insn & 0xff100000) == 0xf9000000) { + /* + * T32 encodings 0b1111_1001_ppp0_qqqq_qqqq_qqqq_qqqq_qqqq + * transform into + * A32 encodings 0b1111_0100_ppp0_qqqq_qqqq_qqqq_qqqq_qqqq + */ + uint32_t a32_insn = (insn & 0x00ffffff) | 0xf4000000; + + if (disas_neon_ls(s, a32_insn)) { + return; + } + } + /* * TODO: Perhaps merge these into one decodetree output file. * Note disas_vfp is written for a32 with cond field in the @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) */ if (disas_t32(s, insn) || disas_vfp_uncond(s, insn) || + disas_neon_shared(s, insn) || ((insn >> 28) == 0xe && disas_vfp(s, insn))) { return; } diff --git a/target/arm/Makefile.objs b/target/arm/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/target/arm/Makefile.objs +++ b/target/arm/Makefile.objs @@ -XXX,XX +XXX,XX @@ target/arm/decode-sve.inc.c: $(SRC_PATH)/target/arm/sve.decode $(DECODETREE) $(PYTHON) $(DECODETREE) --decode disas_sve -o $@ $<,\ "GEN", $(TARGET_DIR)$@) +target/arm/decode-neon-shared.inc.c: $(SRC_PATH)/target/arm/neon-shared.decode $(DECODETREE) + $(call quiet-command,\ + $(PYTHON) $(DECODETREE) --static-decode disas_neon_shared -o $@ $<,\ + "GEN", $(TARGET_DIR)$@) + +target/arm/decode-neon-dp.inc.c: $(SRC_PATH)/target/arm/neon-dp.decode $(DECODETREE) + $(call quiet-command,\ + $(PYTHON) $(DECODETREE) --static-decode disas_neon_dp -o $@ $<,\ + "GEN", $(TARGET_DIR)$@) + +target/arm/decode-neon-ls.inc.c: $(SRC_PATH)/target/arm/neon-ls.decode $(DECODETREE) + $(call quiet-command,\ + $(PYTHON) $(DECODETREE) --static-decode disas_neon_ls -o $@ $<,\ + "GEN", $(TARGET_DIR)$@) + target/arm/decode-vfp.inc.c: $(SRC_PATH)/target/arm/vfp.decode $(DECODETREE) $(call quiet-command,\ $(PYTHON) $(DECODETREE) --static-decode disas_vfp -o $@ $<,\ @@ -XXX,XX +XXX,XX @@ target/arm/decode-t16.inc.c: $(SRC_PATH)/target/arm/t16.decode $(DECODETREE) "GEN", $(TARGET_DIR)$@) target/arm/translate-sve.o: target/arm/decode-sve.inc.c +target/arm/translate.o: target/arm/decode-neon-shared.inc.c +target/arm/translate.o: target/arm/decode-neon-dp.inc.c +target/arm/translate.o: target/arm/decode-neon-ls.inc.c target/arm/translate.o: target/arm/decode-vfp.inc.c target/arm/translate.o: target/arm/decode-vfp-uncond.inc.c target/arm/translate.o: target/arm/decode-a32.inc.c -- 2.20.1 Convert the VCMLA (vector) insns in the 3same extension group to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-5-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 11 ++++++++++ target/arm/translate-neon.inc.c | 37 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 11 +--------- 3 files changed, 49 insertions(+), 10 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ # More specifically, this covers: # 2reg scalar ext: 0b1111_1110_xxxx_xxxx_xxxx_1x0x_xxxx_xxxx # 3same ext: 0b1111_110x_xxxx_xxxx_xxxx_1x0x_xxxx_xxxx + +# VFP/Neon register fields; same as vfp.decode +%vm_dp 5:1 0:4 +%vm_sp 0:4 5:1 +%vn_dp 7:1 16:4 +%vn_sp 16:4 7:1 +%vd_dp 22:1 12:4 +%vd_sp 12:4 22:1 + +VCMLA 1111 110 rot:2 . 1 size:1 .... .... 1000 . q:1 . 0 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ #include "decode-neon-dp.inc.c" #include "decode-neon-ls.inc.c" #include "decode-neon-shared.inc.c" + +static bool trans_VCMLA(DisasContext *s, arg_VCMLA *a) +{ + int opr_sz; + TCGv_ptr fpst; + gen_helper_gvec_3_ptr *fn_gvec_ptr; + + if (!dc_isar_feature(aa32_vcma, s) + || (!a->size && !dc_isar_feature(aa32_fp16_arith, s))) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + ((a->vd | a->vn | a->vm) & 0x10)) { + return false; + } + + if ((a->vn | a->vm | a->vd) & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + opr_sz = (1 + a->q) * 8; + fpst = get_fpstatus_ptr(1); + fn_gvec_ptr = a->size ? gen_helper_gvec_fcmlas : gen_helper_gvec_fcmlah; + tcg_gen_gvec_3_ptr(vfp_reg_offset(1, a->vd), + vfp_reg_offset(1, a->vn), + vfp_reg_offset(1, a->vm), + fpst, opr_sz, opr_sz, a->rot, + fn_gvec_ptr); + tcg_temp_free_ptr(fpst); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_insn_3same_ext(DisasContext *s, uint32_t insn) bool is_long = false, q = extract32(insn, 6, 1); bool ptr_is_env = false; - if ((insn & 0xfe200f10) == 0xfc200800) { - /* VCMLA -- 1111 110R R.1S .... .... 1000 ...0 .... */ - int size = extract32(insn, 20, 1); - data = extract32(insn, 23, 2); /* rot */ - if (!dc_isar_feature(aa32_vcma, s) - || (!size && !dc_isar_feature(aa32_fp16_arith, s))) { - return 1; - } - fn_gvec_ptr = size ? gen_helper_gvec_fcmlas : gen_helper_gvec_fcmlah; - } else if ((insn & 0xfea00f10) == 0xfc800800) { + if ((insn & 0xfea00f10) == 0xfc800800) { /* VCADD -- 1111 110R 1.0S .... .... 1000 ...0 .... */ int size = extract32(insn, 20, 1); data = extract32(insn, 24, 1); /* rot */ -- 2.20.1 Convert the VCADD (vector) insns to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-6-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 3 +++ target/arm/translate-neon.inc.c | 37 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 11 +--------- 3 files changed, 41 insertions(+), 10 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ VCMLA 1111 110 rot:2 . 1 size:1 .... .... 1000 . q:1 . 0 .... \ vm=%vm_dp vn=%vn_dp vd=%vd_dp + +VCADD 1111 110 rot:1 1 . 0 size:1 .... .... 1000 . q:1 . 0 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VCMLA(DisasContext *s, arg_VCMLA *a) tcg_temp_free_ptr(fpst); return true; } + +static bool trans_VCADD(DisasContext *s, arg_VCADD *a) +{ + int opr_sz; + TCGv_ptr fpst; + gen_helper_gvec_3_ptr *fn_gvec_ptr; + + if (!dc_isar_feature(aa32_vcma, s) + || (!a->size && !dc_isar_feature(aa32_fp16_arith, s))) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + ((a->vd | a->vn | a->vm) & 0x10)) { + return false; + } + + if ((a->vn | a->vm | a->vd) & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + opr_sz = (1 + a->q) * 8; + fpst = get_fpstatus_ptr(1); + fn_gvec_ptr = a->size ? gen_helper_gvec_fcadds : gen_helper_gvec_fcaddh; + tcg_gen_gvec_3_ptr(vfp_reg_offset(1, a->vd), + vfp_reg_offset(1, a->vn), + vfp_reg_offset(1, a->vm), + fpst, opr_sz, opr_sz, a->rot, + fn_gvec_ptr); + tcg_temp_free_ptr(fpst); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_insn_3same_ext(DisasContext *s, uint32_t insn) bool is_long = false, q = extract32(insn, 6, 1); bool ptr_is_env = false; - if ((insn & 0xfea00f10) == 0xfc800800) { - /* VCADD -- 1111 110R 1.0S .... .... 1000 ...0 .... */ - int size = extract32(insn, 20, 1); - data = extract32(insn, 24, 1); /* rot */ - if (!dc_isar_feature(aa32_vcma, s) - || (!size && !dc_isar_feature(aa32_fp16_arith, s))) { - return 1; - } - fn_gvec_ptr = size ? gen_helper_gvec_fcadds : gen_helper_gvec_fcaddh; - } else if ((insn & 0xfeb00f00) == 0xfc200d00) { + if ((insn & 0xfeb00f00) == 0xfc200d00) { /* V[US]DOT -- 1111 1100 0.10 .... .... 1101 .Q.U .... */ bool u = extract32(insn, 4, 1); if (!dc_isar_feature(aa32_dp, s)) { -- 2.20.1 Convert the V[US]DOT (vector) insns to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-7-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 4 ++++ target/arm/translate-neon.inc.c | 32 ++++++++++++++++++++++++++++++++ target/arm/translate.c | 9 +-------- 3 files changed, 37 insertions(+), 8 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ VCMLA 1111 110 rot:2 . 1 size:1 .... .... 1000 . q:1 . 0 .... \ VCADD 1111 110 rot:1 1 . 0 size:1 .... .... 1000 . q:1 . 0 .... \ vm=%vm_dp vn=%vn_dp vd=%vd_dp + +# VUDOT and VSDOT +VDOT 1111 110 00 . 10 .... .... 1101 . q:1 . u:1 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VCADD(DisasContext *s, arg_VCADD *a) tcg_temp_free_ptr(fpst); return true; } + +static bool trans_VDOT(DisasContext *s, arg_VDOT *a) +{ + int opr_sz; + gen_helper_gvec_3 *fn_gvec; + + if (!dc_isar_feature(aa32_dp, s)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + ((a->vd | a->vn | a->vm) & 0x10)) { + return false; + } + + if ((a->vn | a->vm | a->vd) & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + opr_sz = (1 + a->q) * 8; + fn_gvec = a->u ? gen_helper_gvec_udot_b : gen_helper_gvec_sdot_b; + tcg_gen_gvec_3_ool(vfp_reg_offset(1, a->vd), + vfp_reg_offset(1, a->vn), + vfp_reg_offset(1, a->vm), + opr_sz, opr_sz, 0, fn_gvec); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_insn_3same_ext(DisasContext *s, uint32_t insn) bool is_long = false, q = extract32(insn, 6, 1); bool ptr_is_env = false; - if ((insn & 0xfeb00f00) == 0xfc200d00) { - /* V[US]DOT -- 1111 1100 0.10 .... .... 1101 .Q.U .... */ - bool u = extract32(insn, 4, 1); - if (!dc_isar_feature(aa32_dp, s)) { - return 1; - } - fn_gvec = u ? gen_helper_gvec_udot_b : gen_helper_gvec_sdot_b; - } else if ((insn & 0xff300f10) == 0xfc200810) { + if ((insn & 0xff300f10) == 0xfc200810) { /* VFM[AS]L -- 1111 1100 S.10 .... .... 1000 .Q.1 .... */ int is_s = extract32(insn, 23, 1); if (!dc_isar_feature(aa32_fhm, s)) { -- 2.20.1 Convert the VFM[AS]L (vector) insns to decodetree. This is the last insn in the legacy decoder for the 3same_ext group, so we can delete the legacy decoder function for the group entirely. Note that in disas_thumb2_insn() the parts of this encoding space where the decodetree decoder returns false will correctly be directed to illegal_op by the "(insn & (1 << 28))" check so they won't fall into disas_coproc_insn() by mistake. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-8-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 6 +++ target/arm/translate-neon.inc.c | 31 +++++++++++ target/arm/translate.c | 92 +-------------------------------- 3 files changed, 38 insertions(+), 91 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ VCADD 1111 110 rot:1 1 . 0 size:1 .... .... 1000 . q:1 . 0 .... \ # VUDOT and VSDOT VDOT 1111 110 00 . 10 .... .... 1101 . q:1 . u:1 .... \ vm=%vm_dp vn=%vn_dp vd=%vd_dp + +# VFM[AS]L +VFML 1111 110 0 s:1 . 10 .... .... 1000 . 0 . 1 .... \ + vm=%vm_sp vn=%vn_sp vd=%vd_dp q=0 +VFML 1111 110 0 s:1 . 10 .... .... 1000 . 1 . 1 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp q=1 diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VDOT(DisasContext *s, arg_VDOT *a) opr_sz, opr_sz, 0, fn_gvec); return true; } + +static bool trans_VFML(DisasContext *s, arg_VFML *a) +{ + int opr_sz; + + if (!dc_isar_feature(aa32_fhm, s)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + (a->vd & 0x10)) { + return false; + } + + if (a->vd & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + opr_sz = (1 + a->q) * 8; + tcg_gen_gvec_3_ptr(vfp_reg_offset(1, a->vd), + vfp_reg_offset(a->q, a->vn), + vfp_reg_offset(a->q, a->vm), + cpu_env, opr_sz, opr_sz, a->s, /* is_2 == 0 */ + gen_helper_gvec_fmlal_a32); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) return 0; } -/* Advanced SIMD three registers of the same length extension. - * 31 25 23 22 20 16 12 11 10 9 8 3 0 - * +---------------+-----+---+-----+----+----+---+----+---+----+---------+----+ - * | 1 1 1 1 1 1 0 | op1 | D | op2 | Vn | Vd | 1 | o3 | 0 | o4 | N Q M U | Vm | - * +---------------+-----+---+-----+----+----+---+----+---+----+---------+----+ - */ -static int disas_neon_insn_3same_ext(DisasContext *s, uint32_t insn) -{ - gen_helper_gvec_3 *fn_gvec = NULL; - gen_helper_gvec_3_ptr *fn_gvec_ptr = NULL; - int rd, rn, rm, opr_sz; - int data = 0; - int off_rn, off_rm; - bool is_long = false, q = extract32(insn, 6, 1); - bool ptr_is_env = false; - - if ((insn & 0xff300f10) == 0xfc200810) { - /* VFM[AS]L -- 1111 1100 S.10 .... .... 1000 .Q.1 .... */ - int is_s = extract32(insn, 23, 1); - if (!dc_isar_feature(aa32_fhm, s)) { - return 1; - } - is_long = true; - data = is_s; /* is_2 == 0 */ - fn_gvec_ptr = gen_helper_gvec_fmlal_a32; - ptr_is_env = true; - } else { - return 1; - } - - VFP_DREG_D(rd, insn); - if (rd & q) { - return 1; - } - if (q || !is_long) { - VFP_DREG_N(rn, insn); - VFP_DREG_M(rm, insn); - if ((rn | rm) & q & !is_long) { - return 1; - } - off_rn = vfp_reg_offset(1, rn); - off_rm = vfp_reg_offset(1, rm); - } else { - rn = VFP_SREG_N(insn); - rm = VFP_SREG_M(insn); - off_rn = vfp_reg_offset(0, rn); - off_rm = vfp_reg_offset(0, rm); - } - - if (s->fp_excp_el) { - gen_exception_insn(s, s->pc_curr, EXCP_UDEF, - syn_simd_access_trap(1, 0xe, false), s->fp_excp_el); - return 0; - } - if (!s->vfp_enabled) { - return 1; - } - - opr_sz = (1 + q) * 8; - if (fn_gvec_ptr) { - TCGv_ptr ptr; - if (ptr_is_env) { - ptr = cpu_env; - } else { - ptr = get_fpstatus_ptr(1); - } - tcg_gen_gvec_3_ptr(vfp_reg_offset(1, rd), off_rn, off_rm, ptr, - opr_sz, opr_sz, data, fn_gvec_ptr); - if (!ptr_is_env) { - tcg_temp_free_ptr(ptr); - } - } else { - tcg_gen_gvec_3_ool(vfp_reg_offset(1, rd), off_rn, off_rm, - opr_sz, opr_sz, data, fn_gvec); - } - return 0; -} - /* Advanced SIMD two registers and a scalar extension. * 31 24 23 22 20 16 12 11 10 9 8 3 0 * +-----------------+----+---+----+----+----+---+----+---+----+---------+----+ @@ -XXX,XX +XXX,XX @@ static void disas_arm_insn(DisasContext *s, unsigned int insn) } } } - } else if ((insn & 0x0e000a00) == 0x0c000800 - && arm_dc_feature(s, ARM_FEATURE_V8)) { - if (disas_neon_insn_3same_ext(s, insn)) { - goto illegal_op; - } - return; } else if ((insn & 0x0f000a00) == 0x0e000800 && arm_dc_feature(s, ARM_FEATURE_V8)) { if (disas_neon_insn_2reg_scalar_ext(s, insn)) { @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) } break; } - if ((insn & 0xfe000a00) == 0xfc000800 + if ((insn & 0xff000a00) == 0xfe000800 && arm_dc_feature(s, ARM_FEATURE_V8)) { /* The Thumb2 and ARM encodings are identical. */ - if (disas_neon_insn_3same_ext(s, insn)) { - goto illegal_op; - } - } else if ((insn & 0xff000a00) == 0xfe000800 - && arm_dc_feature(s, ARM_FEATURE_V8)) { - /* The Thumb2 and ARM encodings are identical. */ if (disas_neon_insn_2reg_scalar_ext(s, insn)) { goto illegal_op; } -- 2.20.1 Convert VCMLA (scalar) in the 2reg-scalar-ext group to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-9-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 5 +++++ target/arm/translate-neon.inc.c | 40 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 26 +-------------------- 3 files changed, 46 insertions(+), 25 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ VFML 1111 110 0 s:1 . 10 .... .... 1000 . 0 . 1 .... \ vm=%vm_sp vn=%vn_sp vd=%vd_dp q=0 VFML 1111 110 0 s:1 . 10 .... .... 1000 . 1 . 1 .... \ vm=%vm_dp vn=%vn_dp vd=%vd_dp q=1 + +VCMLA_scalar 1111 1110 0 . rot:2 .... .... 1000 . q:1 index:1 0 vm:4 \ + vn=%vn_dp vd=%vd_dp size=0 +VCMLA_scalar 1111 1110 1 . rot:2 .... .... 1000 . q:1 . 0 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp size=1 index=0 diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VFML(DisasContext *s, arg_VFML *a) gen_helper_gvec_fmlal_a32); return true; } + +static bool trans_VCMLA_scalar(DisasContext *s, arg_VCMLA_scalar *a) +{ + gen_helper_gvec_3_ptr *fn_gvec_ptr; + int opr_sz; + TCGv_ptr fpst; + + if (!dc_isar_feature(aa32_vcma, s)) { + return false; + } + if (a->size == 0 && !dc_isar_feature(aa32_fp16_arith, s)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + ((a->vd | a->vn | a->vm) & 0x10)) { + return false; + } + + if ((a->vd | a->vn) & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + fn_gvec_ptr = (a->size ? gen_helper_gvec_fcmlas_idx + : gen_helper_gvec_fcmlah_idx); + opr_sz = (1 + a->q) * 8; + fpst = get_fpstatus_ptr(1); + tcg_gen_gvec_3_ptr(vfp_reg_offset(1, a->vd), + vfp_reg_offset(1, a->vn), + vfp_reg_offset(1, a->vm), + fpst, opr_sz, opr_sz, + (a->index << 2) | a->rot, fn_gvec_ptr); + tcg_temp_free_ptr(fpst); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_insn_2reg_scalar_ext(DisasContext *s, uint32_t insn) bool is_long = false, q = extract32(insn, 6, 1); bool ptr_is_env = false; - if ((insn & 0xff000f10) == 0xfe000800) { - /* VCMLA (indexed) -- 1111 1110 S.RR .... .... 1000 ...0 .... */ - int rot = extract32(insn, 20, 2); - int size = extract32(insn, 23, 1); - int index; - - if (!dc_isar_feature(aa32_vcma, s)) { - return 1; - } - if (size == 0) { - if (!dc_isar_feature(aa32_fp16_arith, s)) { - return 1; - } - /* For fp16, rm is just Vm, and index is M. */ - rm = extract32(insn, 0, 4); - index = extract32(insn, 5, 1); - } else { - /* For fp32, rm is the usual M:Vm, and index is 0. */ - VFP_DREG_M(rm, insn); - index = 0; - } - data = (index << 2) | rot; - fn_gvec_ptr = (size ? gen_helper_gvec_fcmlas_idx - : gen_helper_gvec_fcmlah_idx); - } else if ((insn & 0xffb00f00) == 0xfe200d00) { + if ((insn & 0xffb00f00) == 0xfe200d00) { /* V[US]DOT -- 1111 1110 0.10 .... .... 1101 .Q.U .... */ int u = extract32(insn, 4, 1); -- 2.20.1 Convert the V[US]DOT (scalar) insns in the 2reg-scalar-ext group to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-10-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 3 +++ target/arm/translate-neon.inc.c | 35 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 13 +----------- 3 files changed, 39 insertions(+), 12 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ VCMLA_scalar 1111 1110 0 . rot:2 .... .... 1000 . q:1 index:1 0 vm:4 \ vn=%vn_dp vd=%vd_dp size=0 VCMLA_scalar 1111 1110 1 . rot:2 .... .... 1000 . q:1 . 0 .... \ vm=%vm_dp vn=%vn_dp vd=%vd_dp size=1 index=0 + +VDOT_scalar 1111 1110 0 . 10 .... .... 1101 . q:1 index:1 u:1 rm:4 \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VCMLA_scalar(DisasContext *s, arg_VCMLA_scalar *a) tcg_temp_free_ptr(fpst); return true; } + +static bool trans_VDOT_scalar(DisasContext *s, arg_VDOT_scalar *a) +{ + gen_helper_gvec_3 *fn_gvec; + int opr_sz; + TCGv_ptr fpst; + + if (!dc_isar_feature(aa32_dp, s)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + ((a->vd | a->vn) & 0x10)) { + return false; + } + + if ((a->vd | a->vn) & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + fn_gvec = a->u ? gen_helper_gvec_udot_idx_b : gen_helper_gvec_sdot_idx_b; + opr_sz = (1 + a->q) * 8; + fpst = get_fpstatus_ptr(1); + tcg_gen_gvec_3_ool(vfp_reg_offset(1, a->vd), + vfp_reg_offset(1, a->vn), + vfp_reg_offset(1, a->rm), + opr_sz, opr_sz, a->index, fn_gvec); + tcg_temp_free_ptr(fpst); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_insn_2reg_scalar_ext(DisasContext *s, uint32_t insn) bool is_long = false, q = extract32(insn, 6, 1); bool ptr_is_env = false; - if ((insn & 0xffb00f00) == 0xfe200d00) { - /* V[US]DOT -- 1111 1110 0.10 .... .... 1101 .Q.U .... */ - int u = extract32(insn, 4, 1); - - if (!dc_isar_feature(aa32_dp, s)) { - return 1; - } - fn_gvec = u ? gen_helper_gvec_udot_idx_b : gen_helper_gvec_sdot_idx_b; - /* rm is just Vm, and index is M. */ - data = extract32(insn, 5, 1); /* index */ - rm = extract32(insn, 0, 4); - } else if ((insn & 0xffa00f10) == 0xfe000810) { + if ((insn & 0xffa00f10) == 0xfe000810) { /* VFM[AS]L -- 1111 1110 0.0S .... .... 1000 .Q.1 .... */ int is_s = extract32(insn, 20, 1); int vm20 = extract32(insn, 0, 3); -- 2.20.1 Convert the VFM[AS]L (scalar) insns in the 2reg-scalar-ext group to decodetree. These are the last ones in the group so we can remove all the legacy decode for the group. Note that in disas_thumb2_insn() the parts of this encoding space where the decodetree decoder returns false will correctly be directed to illegal_op by the "(insn & (1 << 28))" check so they won't fall into disas_coproc_insn() by mistake. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-11-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 7 +++ target/arm/translate-neon.inc.c | 32 ++++++++++ target/arm/translate.c | 107 +------------------------------- 3 files changed, 40 insertions(+), 106 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ VCMLA_scalar 1111 1110 1 . rot:2 .... .... 1000 . q:1 . 0 .... \ VDOT_scalar 1111 1110 0 . 10 .... .... 1101 . q:1 index:1 u:1 rm:4 \ vm=%vm_dp vn=%vn_dp vd=%vd_dp + +%vfml_scalar_q0_rm 0:3 5:1 +%vfml_scalar_q1_index 5:1 3:1 +VFML_scalar 1111 1110 0 . 0 s:1 .... .... 1000 . 0 . 1 index:1 ... \ + rm=%vfml_scalar_q0_rm vn=%vn_sp vd=%vd_dp q=0 +VFML_scalar 1111 1110 0 . 0 s:1 .... .... 1000 . 1 . 1 . rm:3 \ + index=%vfml_scalar_q1_index vn=%vn_dp vd=%vd_dp q=1 diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VDOT_scalar(DisasContext *s, arg_VDOT_scalar *a) tcg_temp_free_ptr(fpst); return true; } + +static bool trans_VFML_scalar(DisasContext *s, arg_VFML_scalar *a) +{ + int opr_sz; + + if (!dc_isar_feature(aa32_fhm, s)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + ((a->vd & 0x10) || (a->q && (a->vn & 0x10)))) { + return false; + } + + if (a->vd & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + opr_sz = (1 + a->q) * 8; + tcg_gen_gvec_3_ptr(vfp_reg_offset(1, a->vd), + vfp_reg_offset(a->q, a->vn), + vfp_reg_offset(a->q, a->rm), + cpu_env, opr_sz, opr_sz, + (a->index << 2) | a->s, /* is_2 == 0 */ + gen_helper_gvec_fmlal_idx_a32); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_dsp_insn(DisasContext *s, uint32_t insn) } #define VFP_REG_SHR(x, n) (((n) > 0) ? (x) >> (n) : (x) << -(n)) -#define VFP_SREG(insn, bigbit, smallbit) \ - ((VFP_REG_SHR(insn, bigbit - 1) & 0x1e) | (((insn) >> (smallbit)) & 1)) #define VFP_DREG(reg, insn, bigbit, smallbit) do { \ if (dc_isar_feature(aa32_simd_r32, s)) { \ reg = (((insn) >> (bigbit)) & 0x0f) \ @@ -XXX,XX +XXX,XX @@ static int disas_dsp_insn(DisasContext *s, uint32_t insn) reg = ((insn) >> (bigbit)) & 0x0f; \ }} while (0) -#define VFP_SREG_D(insn) VFP_SREG(insn, 12, 22) #define VFP_DREG_D(reg, insn) VFP_DREG(reg, insn, 12, 22) -#define VFP_SREG_N(insn) VFP_SREG(insn, 16, 7) #define VFP_DREG_N(reg, insn) VFP_DREG(reg, insn, 16, 7) -#define VFP_SREG_M(insn) VFP_SREG(insn, 0, 5) #define VFP_DREG_M(reg, insn) VFP_DREG(reg, insn, 0, 5) static void gen_neon_dup_low16(TCGv_i32 var) @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) return 0; } -/* Advanced SIMD two registers and a scalar extension. - * 31 24 23 22 20 16 12 11 10 9 8 3 0 - * +-----------------+----+---+----+----+----+---+----+---+----+---------+----+ - * | 1 1 1 1 1 1 1 0 | o1 | D | o2 | Vn | Vd | 1 | o3 | 0 | o4 | N Q M U | Vm | - * +-----------------+----+---+----+----+----+---+----+---+----+---------+----+ - * - */ - -static int disas_neon_insn_2reg_scalar_ext(DisasContext *s, uint32_t insn) -{ - gen_helper_gvec_3 *fn_gvec = NULL; - gen_helper_gvec_3_ptr *fn_gvec_ptr = NULL; - int rd, rn, rm, opr_sz, data; - int off_rn, off_rm; - bool is_long = false, q = extract32(insn, 6, 1); - bool ptr_is_env = false; - - if ((insn & 0xffa00f10) == 0xfe000810) { - /* VFM[AS]L -- 1111 1110 0.0S .... .... 1000 .Q.1 .... */ - int is_s = extract32(insn, 20, 1); - int vm20 = extract32(insn, 0, 3); - int vm3 = extract32(insn, 3, 1); - int m = extract32(insn, 5, 1); - int index; - - if (!dc_isar_feature(aa32_fhm, s)) { - return 1; - } - if (q) { - rm = vm20; - index = m * 2 + vm3; - } else { - rm = vm20 * 2 + m; - index = vm3; - } - is_long = true; - data = (index << 2) | is_s; /* is_2 == 0 */ - fn_gvec_ptr = gen_helper_gvec_fmlal_idx_a32; - ptr_is_env = true; - } else { - return 1; - } - - VFP_DREG_D(rd, insn); - if (rd & q) { - return 1; - } - if (q || !is_long) { - VFP_DREG_N(rn, insn); - if (rn & q & !is_long) { - return 1; - } - off_rn = vfp_reg_offset(1, rn); - off_rm = vfp_reg_offset(1, rm); - } else { - rn = VFP_SREG_N(insn); - off_rn = vfp_reg_offset(0, rn); - off_rm = vfp_reg_offset(0, rm); - } - if (s->fp_excp_el) { - gen_exception_insn(s, s->pc_curr, EXCP_UDEF, - syn_simd_access_trap(1, 0xe, false), s->fp_excp_el); - return 0; - } - if (!s->vfp_enabled) { - return 1; - } - - opr_sz = (1 + q) * 8; - if (fn_gvec_ptr) { - TCGv_ptr ptr; - if (ptr_is_env) { - ptr = cpu_env; - } else { - ptr = get_fpstatus_ptr(1); - } - tcg_gen_gvec_3_ptr(vfp_reg_offset(1, rd), off_rn, off_rm, ptr, - opr_sz, opr_sz, data, fn_gvec_ptr); - if (!ptr_is_env) { - tcg_temp_free_ptr(ptr); - } - } else { - tcg_gen_gvec_3_ool(vfp_reg_offset(1, rd), off_rn, off_rm, - opr_sz, opr_sz, data, fn_gvec); - } - return 0; -} - static int disas_coproc_insn(DisasContext *s, uint32_t insn) { int cpnum, is64, crn, crm, opc1, opc2, isread, rt, rt2; @@ -XXX,XX +XXX,XX @@ static void disas_arm_insn(DisasContext *s, unsigned int insn) } } } - } else if ((insn & 0x0f000a00) == 0x0e000800 - && arm_dc_feature(s, ARM_FEATURE_V8)) { - if (disas_neon_insn_2reg_scalar_ext(s, insn)) { - goto illegal_op; - } - return; } goto illegal_op; } @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) } break; } - if ((insn & 0xff000a00) == 0xfe000800 - && arm_dc_feature(s, ARM_FEATURE_V8)) { - /* The Thumb2 and ARM encodings are identical. */ - if (disas_neon_insn_2reg_scalar_ext(s, insn)) { - goto illegal_op; - } - } else if (((insn >> 24) & 3) == 3) { + if (((insn >> 24) & 3) == 3) { /* Translate into the equivalent ARM encoding. */ insn = (insn & 0xe2ffffff) | ((insn & (1 << 28)) >> 4) | (1 << 28); if (disas_neon_data_insn(s, insn)) { -- 2.20.1 Convert the Neon "load/store multiple structures" insns to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-12-peter.maydell@linaro.org --- target/arm/neon-ls.decode | 7 ++ target/arm/translate-neon.inc.c | 124 ++++++++++++++++++++++++++++++++ target/arm/translate.c | 91 +---------------------- 3 files changed, 133 insertions(+), 89 deletions(-) diff --git a/target/arm/neon-ls.decode b/target/arm/neon-ls.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-ls.decode +++ b/target/arm/neon-ls.decode @@ -XXX,XX +XXX,XX @@ # 0b1111_1001_xxx0_xxxx_xxxx_xxxx_xxxx_xxxx # This file works on the A32 encoding only; calling code for T32 has to # transform the insn into the A32 version first. + +%vd_dp 22:1 12:4 + +# Neon load/store multiple structures + +VLDST_multiple 1111 0100 0 . l:1 0 rn:4 .... itype:4 size:2 align:2 rm:4 \ + vd=%vd_dp diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VFML_scalar(DisasContext *s, arg_VFML_scalar *a) gen_helper_gvec_fmlal_idx_a32); return true; } + +static struct { + int nregs; + int interleave; + int spacing; +} const neon_ls_element_type[11] = { + {1, 4, 1}, + {1, 4, 2}, + {4, 1, 1}, + {2, 2, 2}, + {1, 3, 1}, + {1, 3, 2}, + {3, 1, 1}, + {1, 1, 1}, + {1, 2, 1}, + {1, 2, 2}, + {2, 1, 1} +}; + +static void gen_neon_ldst_base_update(DisasContext *s, int rm, int rn, + int stride) +{ + if (rm != 15) { + TCGv_i32 base; + + base = load_reg(s, rn); + if (rm == 13) { + tcg_gen_addi_i32(base, base, stride); + } else { + TCGv_i32 index; + index = load_reg(s, rm); + tcg_gen_add_i32(base, base, index); + tcg_temp_free_i32(index); + } + store_reg(s, rn, base); + } +} + +static bool trans_VLDST_multiple(DisasContext *s, arg_VLDST_multiple *a) +{ + /* Neon load/store multiple structures */ + int nregs, interleave, spacing, reg, n; + MemOp endian = s->be_data; + int mmu_idx = get_mem_index(s); + int size = a->size; + TCGv_i64 tmp64; + TCGv_i32 addr, tmp; + + if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist */ + if (!dc_isar_feature(aa32_simd_r32, s) && (a->vd & 0x10)) { + return false; + } + if (a->itype > 10) { + return false; + } + /* Catch UNDEF cases for bad values of align field */ + switch (a->itype & 0xc) { + case 4: + if (a->align >= 2) { + return false; + } + break; + case 8: + if (a->align == 3) { + return false; + } + break; + default: + break; + } + nregs = neon_ls_element_type[a->itype].nregs; + interleave = neon_ls_element_type[a->itype].interleave; + spacing = neon_ls_element_type[a->itype].spacing; + if (size == 3 && (interleave | spacing) != 1) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + /* For our purposes, bytes are always little-endian. */ + if (size == 0) { + endian = MO_LE; + } + /* + * Consecutive little-endian elements from a single register + * can be promoted to a larger little-endian operation. + */ + if (interleave == 1 && endian == MO_LE) { + size = 3; + } + tmp64 = tcg_temp_new_i64(); + addr = tcg_temp_new_i32(); + tmp = tcg_const_i32(1 << size); + load_reg_var(s, addr, a->rn); + for (reg = 0; reg < nregs; reg++) { + for (n = 0; n < 8 >> size; n++) { + int xs; + for (xs = 0; xs < interleave; xs++) { + int tt = a->vd + reg + spacing * xs; + + if (a->l) { + gen_aa32_ld_i64(s, tmp64, addr, mmu_idx, endian | size); + neon_store_element64(tt, n, size, tmp64); + } else { + neon_load_element64(tmp64, tt, n, size); + gen_aa32_st_i64(s, tmp64, addr, mmu_idx, endian | size); + } + tcg_gen_add_i32(addr, addr, tmp); + } + } + } + tcg_temp_free_i32(addr); + tcg_temp_free_i32(tmp); + tcg_temp_free_i64(tmp64); + + gen_neon_ldst_base_update(s, a->rm, a->rn, nregs * interleave * 8); + return true; +} 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 gen_neon_trn_u16(TCGv_i32 t0, TCGv_i32 t1) } -static struct { - int nregs; - int interleave; - int spacing; -} const neon_ls_element_type[11] = { - {1, 4, 1}, - {1, 4, 2}, - {4, 1, 1}, - {2, 2, 2}, - {1, 3, 1}, - {1, 3, 2}, - {3, 1, 1}, - {1, 1, 1}, - {1, 2, 1}, - {1, 2, 2}, - {2, 1, 1} -}; - /* Translate a NEON load/store element instruction. Return nonzero if the instruction is invalid. */ static int disas_neon_ls_insn(DisasContext *s, uint32_t insn) { int rd, rn, rm; - int op; int nregs; - int interleave; - int spacing; int stride; int size; int reg; int load; - int n; int vec_size; - int mmu_idx; - MemOp endian; TCGv_i32 addr; TCGv_i32 tmp; - TCGv_i32 tmp2; - TCGv_i64 tmp64; if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { return 1; @@ -XXX,XX +XXX,XX @@ static int disas_neon_ls_insn(DisasContext *s, uint32_t insn) rn = (insn >> 16) & 0xf; rm = insn & 0xf; load = (insn & (1 << 21)) != 0; - endian = s->be_data; - mmu_idx = get_mem_index(s); if ((insn & (1 << 23)) == 0) { - /* Load store all elements. */ - op = (insn >> 8) & 0xf; - size = (insn >> 6) & 3; - if (op > 10) - return 1; - /* Catch UNDEF cases for bad values of align field */ - switch (op & 0xc) { - case 4: - if (((insn >> 5) & 1) == 1) { - return 1; - } - break; - case 8: - if (((insn >> 4) & 3) == 3) { - return 1; - } - break; - default: - break; - } - nregs = neon_ls_element_type[op].nregs; - interleave = neon_ls_element_type[op].interleave; - spacing = neon_ls_element_type[op].spacing; - if (size == 3 && (interleave | spacing) != 1) { - return 1; - } - /* For our purposes, bytes are always little-endian. */ - if (size == 0) { - endian = MO_LE; - } - /* Consecutive little-endian elements from a single register - * can be promoted to a larger little-endian operation. - */ - if (interleave == 1 && endian == MO_LE) { - size = 3; - } - tmp64 = tcg_temp_new_i64(); - addr = tcg_temp_new_i32(); - tmp2 = tcg_const_i32(1 << size); - load_reg_var(s, addr, rn); - for (reg = 0; reg < nregs; reg++) { - for (n = 0; n < 8 >> size; n++) { - int xs; - for (xs = 0; xs < interleave; xs++) { - int tt = rd + reg + spacing * xs; - - if (load) { - gen_aa32_ld_i64(s, tmp64, addr, mmu_idx, endian | size); - neon_store_element64(tt, n, size, tmp64); - } else { - neon_load_element64(tmp64, tt, n, size); - gen_aa32_st_i64(s, tmp64, addr, mmu_idx, endian | size); - } - tcg_gen_add_i32(addr, addr, tmp2); - } - } - } - tcg_temp_free_i32(addr); - tcg_temp_free_i32(tmp2); - tcg_temp_free_i64(tmp64); - stride = nregs * interleave * 8; + /* Load store all elements -- handled already by decodetree */ + return 1; } else { size = (insn >> 10) & 3; if (size == 3) { -- 2.20.1 Convert the Neon "load single structure to all lanes" insns to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-13-peter.maydell@linaro.org --- target/arm/neon-ls.decode | 5 +++ target/arm/translate-neon.inc.c | 73 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 55 +------------------------ 3 files changed, 80 insertions(+), 53 deletions(-) diff --git a/target/arm/neon-ls.decode b/target/arm/neon-ls.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-ls.decode +++ b/target/arm/neon-ls.decode @@ -XXX,XX +XXX,XX @@ VLDST_multiple 1111 0100 0 . l:1 0 rn:4 .... itype:4 size:2 align:2 rm:4 \ vd=%vd_dp + +# Neon load single element to all lanes + +VLD_all_lanes 1111 0100 1 . 1 0 rn:4 .... 11 n:2 size:2 t:1 a:1 rm:4 \ + vd=%vd_dp diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VLDST_multiple(DisasContext *s, arg_VLDST_multiple *a) gen_neon_ldst_base_update(s, a->rm, a->rn, nregs * interleave * 8); return true; } + +static bool trans_VLD_all_lanes(DisasContext *s, arg_VLD_all_lanes *a) +{ + /* Neon load single structure to all lanes */ + int reg, stride, vec_size; + int vd = a->vd; + int size = a->size; + int nregs = a->n + 1; + TCGv_i32 addr, tmp; + + if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist */ + if (!dc_isar_feature(aa32_simd_r32, s) && (a->vd & 0x10)) { + return false; + } + + if (size == 3) { + if (nregs != 4 || a->a == 0) { + return false; + } + /* For VLD4 size == 3 a == 1 means 32 bits at 16 byte alignment */ + size = 2; + } + if (nregs == 1 && a->a == 1 && size == 0) { + return false; + } + if (nregs == 3 && a->a == 1) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + /* + * VLD1 to all lanes: T bit indicates how many Dregs to write. + * VLD2/3/4 to all lanes: T bit indicates register stride. + */ + stride = a->t ? 2 : 1; + vec_size = nregs == 1 ? stride * 8 : 8; + + tmp = tcg_temp_new_i32(); + addr = tcg_temp_new_i32(); + load_reg_var(s, addr, a->rn); + for (reg = 0; reg < nregs; reg++) { + gen_aa32_ld_i32(s, tmp, addr, get_mem_index(s), + s->be_data | size); + if ((vd & 1) && vec_size == 16) { + /* + * We cannot write 16 bytes at once because the + * destination is unaligned. + */ + tcg_gen_gvec_dup_i32(size, neon_reg_offset(vd, 0), + 8, 8, tmp); + tcg_gen_gvec_mov(0, neon_reg_offset(vd + 1, 0), + neon_reg_offset(vd, 0), 8, 8); + } else { + tcg_gen_gvec_dup_i32(size, neon_reg_offset(vd, 0), + vec_size, vec_size, tmp); + } + tcg_gen_addi_i32(addr, addr, 1 << size); + vd += stride; + } + tcg_temp_free_i32(tmp); + tcg_temp_free_i32(addr); + + gen_neon_ldst_base_update(s, a->rm, a->rn, (1 << size) * nregs); + + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_ls_insn(DisasContext *s, uint32_t insn) int size; int reg; int load; - int vec_size; TCGv_i32 addr; TCGv_i32 tmp; @@ -XXX,XX +XXX,XX @@ static int disas_neon_ls_insn(DisasContext *s, uint32_t insn) } else { size = (insn >> 10) & 3; if (size == 3) { - /* Load single element to all lanes. */ - int a = (insn >> 4) & 1; - if (!load) { - return 1; - } - size = (insn >> 6) & 3; - nregs = ((insn >> 8) & 3) + 1; - - if (size == 3) { - if (nregs != 4 || a == 0) { - return 1; - } - /* For VLD4 size==3 a == 1 means 32 bits at 16 byte alignment */ - size = 2; - } - if (nregs == 1 && a == 1 && size == 0) { - return 1; - } - if (nregs == 3 && a == 1) { - return 1; - } - addr = tcg_temp_new_i32(); - load_reg_var(s, addr, rn); - - /* VLD1 to all lanes: bit 5 indicates how many Dregs to write. - * VLD2/3/4 to all lanes: bit 5 indicates register stride. - */ - stride = (insn & (1 << 5)) ? 2 : 1; - vec_size = nregs == 1 ? stride * 8 : 8; - - tmp = tcg_temp_new_i32(); - for (reg = 0; reg < nregs; reg++) { - gen_aa32_ld_i32(s, tmp, addr, get_mem_index(s), - s->be_data | size); - if ((rd & 1) && vec_size == 16) { - /* We cannot write 16 bytes at once because the - * destination is unaligned. - */ - tcg_gen_gvec_dup_i32(size, neon_reg_offset(rd, 0), - 8, 8, tmp); - tcg_gen_gvec_mov(0, neon_reg_offset(rd + 1, 0), - neon_reg_offset(rd, 0), 8, 8); - } else { - tcg_gen_gvec_dup_i32(size, neon_reg_offset(rd, 0), - vec_size, vec_size, tmp); - } - tcg_gen_addi_i32(addr, addr, 1 << size); - rd += stride; - } - tcg_temp_free_i32(tmp); - tcg_temp_free_i32(addr); - stride = (1 << size) * nregs; + /* Load single element to all lanes -- handled by decodetree */ + return 1; } else { /* Single element. */ int idx = (insn >> 4) & 0xf; -- 2.20.1 Convert the Neon "load/store single structure to one lane" insns to decodetree. As this is the last set of insns in the neon load/store group, we can remove the whole disas_neon_ls_insn() function. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-14-peter.maydell@linaro.org --- target/arm/neon-ls.decode | 11 +++ target/arm/translate-neon.inc.c | 89 +++++++++++++++++++ target/arm/translate.c | 147 -------------------------------- 3 files changed, 100 insertions(+), 147 deletions(-) diff --git a/target/arm/neon-ls.decode b/target/arm/neon-ls.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-ls.decode +++ b/target/arm/neon-ls.decode @@ -XXX,XX +XXX,XX @@ VLDST_multiple 1111 0100 0 . l:1 0 rn:4 .... itype:4 size:2 align:2 rm:4 \ VLD_all_lanes 1111 0100 1 . 1 0 rn:4 .... 11 n:2 size:2 t:1 a:1 rm:4 \ vd=%vd_dp + +# Neon load/store single structure to one lane +%imm1_5_p1 5:1 !function=plus1 +%imm1_6_p1 6:1 !function=plus1 + +VLDST_single 1111 0100 1 . l:1 0 rn:4 .... 00 n:2 reg_idx:3 align:1 rm:4 \ + vd=%vd_dp size=0 stride=1 +VLDST_single 1111 0100 1 . l:1 0 rn:4 .... 01 n:2 reg_idx:2 align:2 rm:4 \ + vd=%vd_dp size=1 stride=%imm1_5_p1 +VLDST_single 1111 0100 1 . l:1 0 rn:4 .... 10 n:2 reg_idx:1 align:3 rm:4 \ + vd=%vd_dp size=2 stride=%imm1_6_p1 diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ * It might be possible to convert it to a standalone .c file eventually. */ +static inline int plus1(DisasContext *s, int x) +{ + return x + 1; +} + /* Include the generated Neon decoder */ #include "decode-neon-dp.inc.c" #include "decode-neon-ls.inc.c" @@ -XXX,XX +XXX,XX @@ static bool trans_VLD_all_lanes(DisasContext *s, arg_VLD_all_lanes *a) return true; } + +static bool trans_VLDST_single(DisasContext *s, arg_VLDST_single *a) +{ + /* Neon load/store single structure to one lane */ + int reg; + int nregs = a->n + 1; + int vd = a->vd; + TCGv_i32 addr, tmp; + + if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist */ + if (!dc_isar_feature(aa32_simd_r32, s) && (a->vd & 0x10)) { + return false; + } + + /* Catch the UNDEF cases. This is unavoidably a bit messy. */ + switch (nregs) { + case 1: + if (((a->align & (1 << a->size)) != 0) || + (a->size == 2 && ((a->align & 3) == 1 || (a->align & 3) == 2))) { + return false; + } + break; + case 3: + if ((a->align & 1) != 0) { + return false; + } + /* fall through */ + case 2: + if (a->size == 2 && (a->align & 2) != 0) { + return false; + } + break; + case 4: + if ((a->size == 2) && ((a->align & 3) == 3)) { + return false; + } + break; + default: + abort(); + } + if ((vd + a->stride * (nregs - 1)) > 31) { + /* + * Attempts to write off the end of the register file are + * UNPREDICTABLE; we choose to UNDEF because otherwise we would + * access off the end of the array that holds the register data. + */ + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + tmp = tcg_temp_new_i32(); + addr = tcg_temp_new_i32(); + load_reg_var(s, addr, a->rn); + /* + * TODO: if we implemented alignment exceptions, we should check + * addr against the alignment encoded in a->align here. + */ + for (reg = 0; reg < nregs; reg++) { + if (a->l) { + gen_aa32_ld_i32(s, tmp, addr, get_mem_index(s), + s->be_data | a->size); + neon_store_element(vd, a->reg_idx, a->size, tmp); + } else { /* Store */ + neon_load_element(tmp, vd, a->reg_idx, a->size); + gen_aa32_st_i32(s, tmp, addr, get_mem_index(s), + s->be_data | a->size); + } + vd += a->stride; + tcg_gen_addi_i32(addr, addr, 1 << a->size); + } + tcg_temp_free_i32(addr); + tcg_temp_free_i32(tmp); + + gen_neon_ldst_base_update(s, a->rm, a->rn, (1 << a->size) * nregs); + + return true; +} 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 gen_neon_trn_u16(TCGv_i32 t0, TCGv_i32 t1) tcg_temp_free_i32(rd); } - -/* Translate a NEON load/store element instruction. Return nonzero if the - instruction is invalid. */ -static int disas_neon_ls_insn(DisasContext *s, uint32_t insn) -{ - int rd, rn, rm; - int nregs; - int stride; - int size; - int reg; - int load; - TCGv_i32 addr; - TCGv_i32 tmp; - - if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { - return 1; - } - - /* FIXME: this access check should not take precedence over UNDEF - * for invalid encodings; we will generate incorrect syndrome information - * for attempts to execute invalid vfp/neon encodings with FP disabled. - */ - if (s->fp_excp_el) { - gen_exception_insn(s, s->pc_curr, EXCP_UDEF, - syn_simd_access_trap(1, 0xe, false), s->fp_excp_el); - return 0; - } - - if (!s->vfp_enabled) - return 1; - VFP_DREG_D(rd, insn); - rn = (insn >> 16) & 0xf; - rm = insn & 0xf; - load = (insn & (1 << 21)) != 0; - if ((insn & (1 << 23)) == 0) { - /* Load store all elements -- handled already by decodetree */ - return 1; - } else { - size = (insn >> 10) & 3; - if (size == 3) { - /* Load single element to all lanes -- handled by decodetree */ - return 1; - } else { - /* Single element. */ - int idx = (insn >> 4) & 0xf; - int reg_idx; - switch (size) { - case 0: - reg_idx = (insn >> 5) & 7; - stride = 1; - break; - case 1: - reg_idx = (insn >> 6) & 3; - stride = (insn & (1 << 5)) ? 2 : 1; - break; - case 2: - reg_idx = (insn >> 7) & 1; - stride = (insn & (1 << 6)) ? 2 : 1; - break; - default: - abort(); - } - nregs = ((insn >> 8) & 3) + 1; - /* Catch the UNDEF cases. This is unavoidably a bit messy. */ - switch (nregs) { - case 1: - if (((idx & (1 << size)) != 0) || - (size == 2 && ((idx & 3) == 1 || (idx & 3) == 2))) { - return 1; - } - break; - case 3: - if ((idx & 1) != 0) { - return 1; - } - /* fall through */ - case 2: - if (size == 2 && (idx & 2) != 0) { - return 1; - } - break; - case 4: - if ((size == 2) && ((idx & 3) == 3)) { - return 1; - } - break; - default: - abort(); - } - if ((rd + stride * (nregs - 1)) > 31) { - /* Attempts to write off the end of the register file - * are UNPREDICTABLE; we choose to UNDEF because otherwise - * the neon_load_reg() would write off the end of the array. - */ - return 1; - } - tmp = tcg_temp_new_i32(); - addr = tcg_temp_new_i32(); - load_reg_var(s, addr, rn); - for (reg = 0; reg < nregs; reg++) { - if (load) { - gen_aa32_ld_i32(s, tmp, addr, get_mem_index(s), - s->be_data | size); - neon_store_element(rd, reg_idx, size, tmp); - } else { /* Store */ - neon_load_element(tmp, rd, reg_idx, size); - gen_aa32_st_i32(s, tmp, addr, get_mem_index(s), - s->be_data | size); - } - rd += stride; - tcg_gen_addi_i32(addr, addr, 1 << size); - } - tcg_temp_free_i32(addr); - tcg_temp_free_i32(tmp); - stride = nregs * (1 << size); - } - } - if (rm != 15) { - TCGv_i32 base; - - base = load_reg(s, rn); - if (rm == 13) { - tcg_gen_addi_i32(base, base, stride); - } else { - TCGv_i32 index; - index = load_reg(s, rm); - tcg_gen_add_i32(base, base, index); - tcg_temp_free_i32(index); - } - store_reg(s, rn, base); - } - return 0; -} - static inline void gen_neon_narrow(int size, TCGv_i32 dest, TCGv_i64 src) { switch (size) { @@ -XXX,XX +XXX,XX @@ static void disas_arm_insn(DisasContext *s, unsigned int insn) } return; } - if ((insn & 0x0f100000) == 0x04000000) { - /* NEON load/store. */ - if (disas_neon_ls_insn(s, insn)) { - goto illegal_op; - } - return; - } if ((insn & 0x0e000f00) == 0x0c000100) { if (arm_dc_feature(s, ARM_FEATURE_IWMMXT)) { /* iWMMXt register transfer. */ @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) } break; case 12: - if ((insn & 0x01100000) == 0x01000000) { - if (disas_neon_ls_insn(s, insn)) { - goto illegal_op; - } - break; - } goto illegal_op; default: illegal_op: -- 2.20.1 Convert the Neon 3-reg-same VADD and VSUB insns to decodetree. Note that we don't need the neon_3r_sizes[op] check here because all size values are OK for VADD and VSUB; we'll add this when we convert the first insn that has size restrictions. For this we need one of the GVecGen*Fn typedefs currently in translate-a64.h; move them all to translate.h as a block so they are visible to the 32-bit decoder. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-15-peter.maydell@linaro.org --- target/arm/translate-a64.h | 9 -------- target/arm/translate.h | 9 ++++++++ target/arm/neon-dp.decode | 17 +++++++++++++++ target/arm/translate-neon.inc.c | 38 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 14 ++++-------- 5 files changed, 68 insertions(+), 19 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 @@ static inline int vec_full_reg_size(DisasContext *s) bool disas_sve(DisasContext *, uint32_t); -/* Note that the gvec expanders operate on offsets + sizes. */ -typedef void GVecGen2Fn(unsigned, uint32_t, uint32_t, uint32_t, uint32_t); -typedef void GVecGen2iFn(unsigned, uint32_t, uint32_t, int64_t, - uint32_t, uint32_t); -typedef void GVecGen3Fn(unsigned, uint32_t, uint32_t, - uint32_t, uint32_t, uint32_t); -typedef void GVecGen4Fn(unsigned, uint32_t, uint32_t, uint32_t, - uint32_t, uint32_t, uint32_t); - #endif /* TARGET_ARM_TRANSLATE_A64_H */ 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 @@ void gen_sshl_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b); #define dc_isar_feature(name, ctx) \ ({ DisasContext *ctx_ = (ctx); isar_feature_##name(ctx_->isar); }) +/* Note that the gvec expanders operate on offsets + sizes. */ +typedef void GVecGen2Fn(unsigned, uint32_t, uint32_t, uint32_t, uint32_t); +typedef void GVecGen2iFn(unsigned, uint32_t, uint32_t, int64_t, + uint32_t, uint32_t); +typedef void GVecGen3Fn(unsigned, uint32_t, uint32_t, + uint32_t, uint32_t, uint32_t); +typedef void GVecGen4Fn(unsigned, uint32_t, uint32_t, uint32_t, + uint32_t, uint32_t, uint32_t); + #endif /* TARGET_ARM_TRANSLATE_H */ diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-dp.decode +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ # # This file is processed by scripts/decodetree.py # +# VFP/Neon register fields; same as vfp.decode +%vm_dp 5:1 0:4 +%vn_dp 7:1 16:4 +%vd_dp 22:1 12:4 # Encodings for Neon data processing instructions where the T32 encoding # is a simple transformation of the A32 encoding. @@ -XXX,XX +XXX,XX @@ # 0b111p_1111_qqqq_qqqq_qqqq_qqqq_qqqq_qqqq # This file works on the A32 encoding only; calling code for T32 has to # transform the insn into the A32 version first. + +###################################################################### +# 3-reg-same grouping: +# 1111 001 U 0 D sz:2 Vn:4 Vd:4 opc:4 N Q M op Vm:4 +###################################################################### + +&3same vm vn vd q size + +@3same .... ... . . . size:2 .... .... .... . q:1 . . .... \ + &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp + +VADD_3s 1111 001 0 0 . .. .... .... 1000 . . . 0 .... @3same +VSUB_3s 1111 001 1 0 . .. .... .... 1000 . . . 0 .... @3same diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VLDST_single(DisasContext *s, arg_VLDST_single *a) return true; } + +static bool do_3same(DisasContext *s, arg_3same *a, GVecGen3Fn fn) +{ + int vec_size = a->q ? 16 : 8; + int rd_ofs = neon_reg_offset(a->vd, 0); + int rn_ofs = neon_reg_offset(a->vn, 0); + int rm_ofs = neon_reg_offset(a->vm, 0); + + if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + ((a->vd | a->vn | a->vm) & 0x10)) { + return false; + } + + if ((a->vn | a->vm | a->vd) & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + fn(a->size, rd_ofs, rn_ofs, rm_ofs, vec_size, vec_size); + return true; +} + +#define DO_3SAME(INSN, FUNC) \ + static bool trans_##INSN##_3s(DisasContext *s, arg_3same *a) \ + { \ + return do_3same(s, a, FUNC); \ + } + +DO_3SAME(VADD, tcg_gen_gvec_add) +DO_3SAME(VSUB, tcg_gen_gvec_sub) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) } return 0; - case NEON_3R_VADD_VSUB: - if (u) { - tcg_gen_gvec_sub(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } else { - tcg_gen_gvec_add(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } - return 0; - case NEON_3R_VQADD: tcg_gen_gvec_4(rd_ofs, offsetof(CPUARMState, vfp.qc), rn_ofs, rm_ofs, vec_size, vec_size, @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) tcg_gen_gvec_3(rd_ofs, rm_ofs, rn_ofs, vec_size, vec_size, u ? &ushl_op[size] : &sshl_op[size]); return 0; + + case NEON_3R_VADD_VSUB: + /* Already handled by decodetree */ + return 1; } if (size == 3) { -- 2.20.1 Convert the Neon logic ops in the 3-reg-same grouping to decodetree. Note that for the logic ops the 'size' field forms part of their decode and the actual operations are always bitwise. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-16-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 12 +++++++++++ target/arm/translate-neon.inc.c | 19 +++++++++++++++++ target/arm/translate.c | 38 +-------------------------------- 3 files changed, 32 insertions(+), 37 deletions(-) diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-dp.decode +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ @3same .... ... . . . size:2 .... .... .... . q:1 . . .... \ &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp +@3same_logic .... ... . . . .. .... .... .... . q:1 .. .... \ + &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp size=0 + +VAND_3s 1111 001 0 0 . 00 .... .... 0001 ... 1 .... @3same_logic +VBIC_3s 1111 001 0 0 . 01 .... .... 0001 ... 1 .... @3same_logic +VORR_3s 1111 001 0 0 . 10 .... .... 0001 ... 1 .... @3same_logic +VORN_3s 1111 001 0 0 . 11 .... .... 0001 ... 1 .... @3same_logic +VEOR_3s 1111 001 1 0 . 00 .... .... 0001 ... 1 .... @3same_logic +VBSL_3s 1111 001 1 0 . 01 .... .... 0001 ... 1 .... @3same_logic +VBIT_3s 1111 001 1 0 . 10 .... .... 0001 ... 1 .... @3same_logic +VBIF_3s 1111 001 1 0 . 11 .... .... 0001 ... 1 .... @3same_logic + VADD_3s 1111 001 0 0 . .. .... .... 1000 . . . 0 .... @3same VSUB_3s 1111 001 1 0 . .. .... .... 1000 . . . 0 .... @3same diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool do_3same(DisasContext *s, arg_3same *a, GVecGen3Fn fn) DO_3SAME(VADD, tcg_gen_gvec_add) DO_3SAME(VSUB, tcg_gen_gvec_sub) +DO_3SAME(VAND, tcg_gen_gvec_and) +DO_3SAME(VBIC, tcg_gen_gvec_andc) +DO_3SAME(VORR, tcg_gen_gvec_or) +DO_3SAME(VORN, tcg_gen_gvec_orc) +DO_3SAME(VEOR, tcg_gen_gvec_xor) + +/* These insns are all gvec_bitsel but with the inputs in various orders. */ +#define DO_3SAME_BITSEL(INSN, O1, O2, O3) \ + static void gen_##INSN##_3s(unsigned vece, uint32_t rd_ofs, \ + uint32_t rn_ofs, uint32_t rm_ofs, \ + uint32_t oprsz, uint32_t maxsz) \ + { \ + tcg_gen_gvec_bitsel(vece, rd_ofs, O1, O2, O3, oprsz, maxsz); \ + } \ + DO_3SAME(INSN, gen_##INSN##_3s) + +DO_3SAME_BITSEL(VBSL, rd_ofs, rn_ofs, rm_ofs) +DO_3SAME_BITSEL(VBIT, rm_ofs, rn_ofs, rd_ofs) +DO_3SAME_BITSEL(VBIF, rm_ofs, rd_ofs, rn_ofs) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) } return 1; - case NEON_3R_LOGIC: /* Logic ops. */ - switch ((u << 2) | size) { - case 0: /* VAND */ - tcg_gen_gvec_and(0, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - break; - case 1: /* VBIC */ - tcg_gen_gvec_andc(0, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - break; - case 2: /* VORR */ - tcg_gen_gvec_or(0, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - break; - case 3: /* VORN */ - tcg_gen_gvec_orc(0, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - break; - case 4: /* VEOR */ - tcg_gen_gvec_xor(0, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - break; - case 5: /* VBSL */ - tcg_gen_gvec_bitsel(MO_8, rd_ofs, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - break; - case 6: /* VBIT */ - tcg_gen_gvec_bitsel(MO_8, rd_ofs, rm_ofs, rn_ofs, rd_ofs, - vec_size, vec_size); - break; - case 7: /* VBIF */ - tcg_gen_gvec_bitsel(MO_8, rd_ofs, rm_ofs, rd_ofs, rn_ofs, - vec_size, vec_size); - break; - } - return 0; - case NEON_3R_VQADD: tcg_gen_gvec_4(rd_ofs, offsetof(CPUARMState, vfp.qc), rn_ofs, rm_ofs, vec_size, vec_size, @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) return 0; case NEON_3R_VADD_VSUB: + case NEON_3R_LOGIC: /* Already handled by decodetree */ return 1; } -- 2.20.1 Convert the Neon 3-reg-same VMAX and VMIN insns to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-17-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 5 +++++ target/arm/translate-neon.inc.c | 14 ++++++++++++++ target/arm/translate.c | 21 ++------------------- 3 files changed, 21 insertions(+), 19 deletions(-) diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-dp.decode +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ VBSL_3s 1111 001 1 0 . 01 .... .... 0001 ... 1 .... @3same_logic VBIT_3s 1111 001 1 0 . 10 .... .... 0001 ... 1 .... @3same_logic VBIF_3s 1111 001 1 0 . 11 .... .... 0001 ... 1 .... @3same_logic +VMAX_S_3s 1111 001 0 0 . .. .... .... 0110 . . . 0 .... @3same +VMAX_U_3s 1111 001 1 0 . .. .... .... 0110 . . . 0 .... @3same +VMIN_S_3s 1111 001 0 0 . .. .... .... 0110 . . . 1 .... @3same +VMIN_U_3s 1111 001 1 0 . .. .... .... 0110 . . . 1 .... @3same + VADD_3s 1111 001 0 0 . .. .... .... 1000 . . . 0 .... @3same VSUB_3s 1111 001 1 0 . .. .... .... 1000 . . . 0 .... @3same diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ DO_3SAME(VEOR, tcg_gen_gvec_xor) DO_3SAME_BITSEL(VBSL, rd_ofs, rn_ofs, rm_ofs) DO_3SAME_BITSEL(VBIT, rm_ofs, rn_ofs, rd_ofs) DO_3SAME_BITSEL(VBIF, rm_ofs, rd_ofs, rn_ofs) + +#define DO_3SAME_NO_SZ_3(INSN, FUNC) \ + static bool trans_##INSN##_3s(DisasContext *s, arg_3same *a) \ + { \ + if (a->size == 3) { \ + return false; \ + } \ + return do_3same(s, a, FUNC); \ + } + +DO_3SAME_NO_SZ_3(VMAX_S, tcg_gen_gvec_smax) +DO_3SAME_NO_SZ_3(VMAX_U, tcg_gen_gvec_umax) +DO_3SAME_NO_SZ_3(VMIN_S, tcg_gen_gvec_smin) +DO_3SAME_NO_SZ_3(VMIN_U, tcg_gen_gvec_umin) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) rd_ofs, rn_ofs, rm_ofs, vec_size, vec_size); return 0; - case NEON_3R_VMAX: - if (u) { - tcg_gen_gvec_umax(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } else { - tcg_gen_gvec_smax(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } - return 0; - case NEON_3R_VMIN: - if (u) { - tcg_gen_gvec_umin(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } else { - tcg_gen_gvec_smin(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } - return 0; - case NEON_3R_VSHL: /* Note the operation is vshl vd,vm,vn */ tcg_gen_gvec_3(rd_ofs, rm_ofs, rn_ofs, vec_size, vec_size, @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) case NEON_3R_VADD_VSUB: case NEON_3R_LOGIC: + case NEON_3R_VMAX: + case NEON_3R_VMIN: /* Already handled by decodetree */ return 1; } -- 2.20.1 Convert the Neon comparison ops in the 3-reg-same grouping to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-18-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 8 ++++++++ target/arm/translate-neon.inc.c | 22 ++++++++++++++++++++++ target/arm/translate.c | 23 +++-------------------- 3 files changed, 33 insertions(+), 20 deletions(-) diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-dp.decode +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ VBSL_3s 1111 001 1 0 . 01 .... .... 0001 ... 1 .... @3same_logic VBIT_3s 1111 001 1 0 . 10 .... .... 0001 ... 1 .... @3same_logic VBIF_3s 1111 001 1 0 . 11 .... .... 0001 ... 1 .... @3same_logic +VCGT_S_3s 1111 001 0 0 . .. .... .... 0011 . . . 0 .... @3same +VCGT_U_3s 1111 001 1 0 . .. .... .... 0011 . . . 0 .... @3same +VCGE_S_3s 1111 001 0 0 . .. .... .... 0011 . . . 1 .... @3same +VCGE_U_3s 1111 001 1 0 . .. .... .... 0011 . . . 1 .... @3same + VMAX_S_3s 1111 001 0 0 . .. .... .... 0110 . . . 0 .... @3same VMAX_U_3s 1111 001 1 0 . .. .... .... 0110 . . . 0 .... @3same VMIN_S_3s 1111 001 0 0 . .. .... .... 0110 . . . 1 .... @3same @@ -XXX,XX +XXX,XX @@ VMIN_U_3s 1111 001 1 0 . .. .... .... 0110 . . . 1 .... @3same VADD_3s 1111 001 0 0 . .. .... .... 1000 . . . 0 .... @3same VSUB_3s 1111 001 1 0 . .. .... .... 1000 . . . 0 .... @3same + +VTST_3s 1111 001 0 0 . .. .... .... 1000 . . . 1 .... @3same +VCEQ_3s 1111 001 1 0 . .. .... .... 1000 . . . 1 .... @3same diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ DO_3SAME_NO_SZ_3(VMAX_S, tcg_gen_gvec_smax) DO_3SAME_NO_SZ_3(VMAX_U, tcg_gen_gvec_umax) DO_3SAME_NO_SZ_3(VMIN_S, tcg_gen_gvec_smin) DO_3SAME_NO_SZ_3(VMIN_U, tcg_gen_gvec_umin) + +#define DO_3SAME_CMP(INSN, COND) \ + static void gen_##INSN##_3s(unsigned vece, uint32_t rd_ofs, \ + uint32_t rn_ofs, uint32_t rm_ofs, \ + uint32_t oprsz, uint32_t maxsz) \ + { \ + tcg_gen_gvec_cmp(COND, vece, rd_ofs, rn_ofs, rm_ofs, oprsz, maxsz); \ + } \ + DO_3SAME_NO_SZ_3(INSN, gen_##INSN##_3s) + +DO_3SAME_CMP(VCGT_S, TCG_COND_GT) +DO_3SAME_CMP(VCGT_U, TCG_COND_GTU) +DO_3SAME_CMP(VCGE_S, TCG_COND_GE) +DO_3SAME_CMP(VCGE_U, TCG_COND_GEU) +DO_3SAME_CMP(VCEQ, TCG_COND_EQ) + +static void gen_VTST_3s(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t oprsz, uint32_t maxsz) +{ + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, oprsz, maxsz, &cmtst_op[vece]); +} +DO_3SAME_NO_SZ_3(VTST, gen_VTST_3s) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) u ? &mls_op[size] : &mla_op[size]); return 0; - case NEON_3R_VTST_VCEQ: - if (u) { /* VCEQ */ - tcg_gen_gvec_cmp(TCG_COND_EQ, size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } else { /* VTST */ - tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size, &cmtst_op[size]); - } - return 0; - - case NEON_3R_VCGT: - tcg_gen_gvec_cmp(u ? TCG_COND_GTU : TCG_COND_GT, size, - rd_ofs, rn_ofs, rm_ofs, vec_size, vec_size); - return 0; - - case NEON_3R_VCGE: - tcg_gen_gvec_cmp(u ? TCG_COND_GEU : TCG_COND_GE, size, - rd_ofs, rn_ofs, rm_ofs, vec_size, vec_size); - return 0; - case NEON_3R_VSHL: /* Note the operation is vshl vd,vm,vn */ tcg_gen_gvec_3(rd_ofs, rm_ofs, rn_ofs, vec_size, vec_size, @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) case NEON_3R_LOGIC: case NEON_3R_VMAX: case NEON_3R_VMIN: + case NEON_3R_VTST_VCEQ: + case NEON_3R_VCGT: + case NEON_3R_VCGE: /* Already handled by decodetree */ return 1; } -- 2.20.1 Convert the Neon VQADD/VQSUB insns in the 3-reg-same grouping to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-19-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 6 ++++++ target/arm/translate-neon.inc.c | 15 +++++++++++++++ target/arm/translate.c | 14 ++------------ 3 files changed, 23 insertions(+), 12 deletions(-) diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-dp.decode +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ @3same .... ... . . . size:2 .... .... .... . q:1 . . .... \ &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp +VQADD_S_3s 1111 001 0 0 . .. .... .... 0000 . . . 1 .... @3same +VQADD_U_3s 1111 001 1 0 . .. .... .... 0000 . . . 1 .... @3same + @3same_logic .... ... . . . .. .... .... .... . q:1 .. .... \ &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp size=0 @@ -XXX,XX +XXX,XX @@ VBSL_3s 1111 001 1 0 . 01 .... .... 0001 ... 1 .... @3same_logic VBIT_3s 1111 001 1 0 . 10 .... .... 0001 ... 1 .... @3same_logic VBIF_3s 1111 001 1 0 . 11 .... .... 0001 ... 1 .... @3same_logic +VQSUB_S_3s 1111 001 0 0 . .. .... .... 0010 . . . 1 .... @3same +VQSUB_U_3s 1111 001 1 0 . .. .... .... 0010 . . . 1 .... @3same + VCGT_S_3s 1111 001 0 0 . .. .... .... 0011 . . . 0 .... @3same VCGT_U_3s 1111 001 1 0 . .. .... .... 0011 . . . 0 .... @3same VCGE_S_3s 1111 001 0 0 . .. .... .... 0011 . . . 1 .... @3same diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static void gen_VTST_3s(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, oprsz, maxsz, &cmtst_op[vece]); } DO_3SAME_NO_SZ_3(VTST, gen_VTST_3s) + +#define DO_3SAME_GVEC4(INSN, OPARRAY) \ + static void gen_##INSN##_3s(unsigned vece, uint32_t rd_ofs, \ + uint32_t rn_ofs, uint32_t rm_ofs, \ + uint32_t oprsz, uint32_t maxsz) \ + { \ + tcg_gen_gvec_4(rd_ofs, offsetof(CPUARMState, vfp.qc), \ + rn_ofs, rm_ofs, oprsz, maxsz, &OPARRAY[vece]); \ + } \ + DO_3SAME(INSN, gen_##INSN##_3s) + +DO_3SAME_GVEC4(VQADD_S, sqadd_op) +DO_3SAME_GVEC4(VQADD_U, uqadd_op) +DO_3SAME_GVEC4(VQSUB_S, sqsub_op) +DO_3SAME_GVEC4(VQSUB_U, uqsub_op) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) } return 1; - case NEON_3R_VQADD: - tcg_gen_gvec_4(rd_ofs, offsetof(CPUARMState, vfp.qc), - rn_ofs, rm_ofs, vec_size, vec_size, - (u ? uqadd_op : sqadd_op) + size); - return 0; - - case NEON_3R_VQSUB: - tcg_gen_gvec_4(rd_ofs, offsetof(CPUARMState, vfp.qc), - rn_ofs, rm_ofs, vec_size, vec_size, - (u ? uqsub_op : sqsub_op) + size); - return 0; - case NEON_3R_VMUL: /* VMUL */ if (u) { /* Polynomial case allows only P8. */ @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) case NEON_3R_VTST_VCEQ: case NEON_3R_VCGT: case NEON_3R_VCGE: + case NEON_3R_VQADD: + case NEON_3R_VQSUB: /* Already handled by decodetree */ return 1; } -- 2.20.1 Convert the Neon VMUL, VMLA, VMLS and VSHL insns in the 3-reg-same grouping to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-20-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 9 +++++++ target/arm/translate-neon.inc.c | 44 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 28 +++------------------ 3 files changed, 56 insertions(+), 25 deletions(-) diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-dp.decode +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ VCGT_U_3s 1111 001 1 0 . .. .... .... 0011 . . . 0 .... @3same VCGE_S_3s 1111 001 0 0 . .. .... .... 0011 . . . 1 .... @3same VCGE_U_3s 1111 001 1 0 . .. .... .... 0011 . . . 1 .... @3same +VSHL_S_3s 1111 001 0 0 . .. .... .... 0100 . . . 0 .... @3same +VSHL_U_3s 1111 001 1 0 . .. .... .... 0100 . . . 0 .... @3same + VMAX_S_3s 1111 001 0 0 . .. .... .... 0110 . . . 0 .... @3same VMAX_U_3s 1111 001 1 0 . .. .... .... 0110 . . . 0 .... @3same VMIN_S_3s 1111 001 0 0 . .. .... .... 0110 . . . 1 .... @3same @@ -XXX,XX +XXX,XX @@ VSUB_3s 1111 001 1 0 . .. .... .... 1000 . . . 0 .... @3same VTST_3s 1111 001 0 0 . .. .... .... 1000 . . . 1 .... @3same VCEQ_3s 1111 001 1 0 . .. .... .... 1000 . . . 1 .... @3same + +VMLA_3s 1111 001 0 0 . .. .... .... 1001 . . . 0 .... @3same +VMLS_3s 1111 001 1 0 . .. .... .... 1001 . . . 0 .... @3same + +VMUL_3s 1111 001 0 0 . .. .... .... 1001 . . . 1 .... @3same +VMUL_p_3s 1111 001 1 0 . .. .... .... 1001 . . . 1 .... @3same diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ DO_3SAME_NO_SZ_3(VMAX_S, tcg_gen_gvec_smax) DO_3SAME_NO_SZ_3(VMAX_U, tcg_gen_gvec_umax) DO_3SAME_NO_SZ_3(VMIN_S, tcg_gen_gvec_smin) DO_3SAME_NO_SZ_3(VMIN_U, tcg_gen_gvec_umin) +DO_3SAME_NO_SZ_3(VMUL, tcg_gen_gvec_mul) #define DO_3SAME_CMP(INSN, COND) \ static void gen_##INSN##_3s(unsigned vece, uint32_t rd_ofs, \ @@ -XXX,XX +XXX,XX @@ DO_3SAME_GVEC4(VQADD_S, sqadd_op) DO_3SAME_GVEC4(VQADD_U, uqadd_op) DO_3SAME_GVEC4(VQSUB_S, sqsub_op) DO_3SAME_GVEC4(VQSUB_U, uqsub_op) + +static void gen_VMUL_p_3s(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t oprsz, uint32_t maxsz) +{ + tcg_gen_gvec_3_ool(rd_ofs, rn_ofs, rm_ofs, oprsz, maxsz, + 0, gen_helper_gvec_pmul_b); +} + +static bool trans_VMUL_p_3s(DisasContext *s, arg_3same *a) +{ + if (a->size != 0) { + return false; + } + return do_3same(s, a, gen_VMUL_p_3s); +} + +#define DO_3SAME_GVEC3_NO_SZ_3(INSN, OPARRAY) \ + static void gen_##INSN##_3s(unsigned vece, uint32_t rd_ofs, \ + uint32_t rn_ofs, uint32_t rm_ofs, \ + uint32_t oprsz, uint32_t maxsz) \ + { \ + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, \ + oprsz, maxsz, &OPARRAY[vece]); \ + } \ + DO_3SAME_NO_SZ_3(INSN, gen_##INSN##_3s) + + +DO_3SAME_GVEC3_NO_SZ_3(VMLA, mla_op) +DO_3SAME_GVEC3_NO_SZ_3(VMLS, mls_op) + +#define DO_3SAME_GVEC3_SHIFT(INSN, OPARRAY) \ + static void gen_##INSN##_3s(unsigned vece, uint32_t rd_ofs, \ + uint32_t rn_ofs, uint32_t rm_ofs, \ + uint32_t oprsz, uint32_t maxsz) \ + { \ + /* Note the operation is vshl vd,vm,vn */ \ + tcg_gen_gvec_3(rd_ofs, rm_ofs, rn_ofs, \ + oprsz, maxsz, &OPARRAY[vece]); \ + } \ + DO_3SAME(INSN, gen_##INSN##_3s) + +DO_3SAME_GVEC3_SHIFT(VSHL_S, sshl_op) +DO_3SAME_GVEC3_SHIFT(VSHL_U, ushl_op) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) } return 1; - case NEON_3R_VMUL: /* VMUL */ - if (u) { - /* Polynomial case allows only P8. */ - if (size != 0) { - return 1; - } - tcg_gen_gvec_3_ool(rd_ofs, rn_ofs, rm_ofs, vec_size, vec_size, - 0, gen_helper_gvec_pmul_b); - } else { - tcg_gen_gvec_mul(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } - return 0; - - case NEON_3R_VML: /* VMLA, VMLS */ - tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, vec_size, vec_size, - u ? &mls_op[size] : &mla_op[size]); - return 0; - - case NEON_3R_VSHL: - /* Note the operation is vshl vd,vm,vn */ - tcg_gen_gvec_3(rd_ofs, rm_ofs, rn_ofs, vec_size, vec_size, - u ? &ushl_op[size] : &sshl_op[size]); - return 0; - case NEON_3R_VADD_VSUB: case NEON_3R_LOGIC: case NEON_3R_VMAX: @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) case NEON_3R_VCGE: case NEON_3R_VQADD: case NEON_3R_VQSUB: + case NEON_3R_VMUL: + case NEON_3R_VML: + case NEON_3R_VSHL: /* Already handled by decodetree */ return 1; } -- 2.20.1 We're going to want at least some of the NeonGen* typedefs for the refactored 32-bit Neon decoder, so move them all to translate.h since it makes more sense to keep them in one group. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-23-peter.maydell@linaro.org --- target/arm/translate.h | 17 +++++++++++++++++ target/arm/translate-a64.c | 17 ----------------- 2 files changed, 17 insertions(+), 17 deletions(-) diff --git a/target/arm/translate.h b/target/arm/translate.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.h +++ b/target/arm/translate.h @@ -XXX,XX +XXX,XX @@ typedef void GVecGen3Fn(unsigned, uint32_t, uint32_t, typedef void GVecGen4Fn(unsigned, uint32_t, uint32_t, uint32_t, uint32_t, uint32_t, uint32_t); +/* Function prototype for gen_ functions for calling Neon helpers */ +typedef void NeonGenOneOpEnvFn(TCGv_i32, TCGv_ptr, TCGv_i32); +typedef void NeonGenTwoOpFn(TCGv_i32, TCGv_i32, TCGv_i32); +typedef void NeonGenTwoOpEnvFn(TCGv_i32, TCGv_ptr, TCGv_i32, TCGv_i32); +typedef void NeonGenTwo64OpFn(TCGv_i64, TCGv_i64, TCGv_i64); +typedef void NeonGenTwo64OpEnvFn(TCGv_i64, TCGv_ptr, TCGv_i64, TCGv_i64); +typedef void NeonGenNarrowFn(TCGv_i32, TCGv_i64); +typedef void NeonGenNarrowEnvFn(TCGv_i32, TCGv_ptr, TCGv_i64); +typedef void NeonGenWidenFn(TCGv_i64, TCGv_i32); +typedef void NeonGenTwoSingleOPFn(TCGv_i32, TCGv_i32, TCGv_i32, TCGv_ptr); +typedef void NeonGenTwoDoubleOPFn(TCGv_i64, TCGv_i64, TCGv_i64, TCGv_ptr); +typedef void NeonGenOneOpFn(TCGv_i64, TCGv_i64); +typedef void CryptoTwoOpFn(TCGv_ptr, TCGv_ptr); +typedef void CryptoThreeOpIntFn(TCGv_ptr, TCGv_ptr, TCGv_i32); +typedef void CryptoThreeOpFn(TCGv_ptr, TCGv_ptr, TCGv_ptr); +typedef void AtomicThreeOpFn(TCGv_i64, TCGv_i64, TCGv_i64, TCGArg, MemOp); + #endif /* TARGET_ARM_TRANSLATE_H */ diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ typedef struct AArch64DecodeTable { AArch64DecodeFn *disas_fn; } AArch64DecodeTable; -/* Function prototype for gen_ functions for calling Neon helpers */ -typedef void NeonGenOneOpEnvFn(TCGv_i32, TCGv_ptr, TCGv_i32); -typedef void NeonGenTwoOpFn(TCGv_i32, TCGv_i32, TCGv_i32); -typedef void NeonGenTwoOpEnvFn(TCGv_i32, TCGv_ptr, TCGv_i32, TCGv_i32); -typedef void NeonGenTwo64OpFn(TCGv_i64, TCGv_i64, TCGv_i64); -typedef void NeonGenTwo64OpEnvFn(TCGv_i64, TCGv_ptr, TCGv_i64, TCGv_i64); -typedef void NeonGenNarrowFn(TCGv_i32, TCGv_i64); -typedef void NeonGenNarrowEnvFn(TCGv_i32, TCGv_ptr, TCGv_i64); -typedef void NeonGenWidenFn(TCGv_i64, TCGv_i32); -typedef void NeonGenTwoSingleOPFn(TCGv_i32, TCGv_i32, TCGv_i32, TCGv_ptr); -typedef void NeonGenTwoDoubleOPFn(TCGv_i64, TCGv_i64, TCGv_i64, TCGv_ptr); -typedef void NeonGenOneOpFn(TCGv_i64, TCGv_i64); -typedef void CryptoTwoOpFn(TCGv_ptr, TCGv_ptr); -typedef void CryptoThreeOpIntFn(TCGv_ptr, TCGv_ptr, TCGv_i32); -typedef void CryptoThreeOpFn(TCGv_ptr, TCGv_ptr, TCGv_ptr); -typedef void AtomicThreeOpFn(TCGv_i64, TCGv_i64, TCGv_i64, TCGArg, MemOp); - /* initialize TCG globals. */ void a64_translate_init(void) { -- 2.20.1 Mostly this is RTH's memtag series, but there are also some cleanups from Philippe. thanks -- PMM The following changes since commit 10f7ffabf9c507fc02382b89912003b1c43c3231: Merge remote-tracking branch 'remotes/mcayland/tags/qemu-macppc-20200626' into staging (2020-06-26 12:14:18 +0100) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20200626 for you to fetch changes up to c7459633baa71d1781fde4a245d6ec9ce2f008cf: target/arm: Enable MTE (2020-06-26 14:32:24 +0100) ---------------------------------------------------------------- target-arm queue: * hw/arm/aspeed: improve QOM usage * hw/misc/pca9552: trace GPIO change events * target/arm: Implement ARMv8.5-MemTag for system emulation ---------------------------------------------------------------- Philippe Mathieu-Daudé (12): hw/arm/aspeed: Remove extraneous MemoryRegion object owner hw/arm/aspeed: Rename AspeedBoardState as AspeedMachineState hw/arm/aspeed: QOM'ify AspeedMachineState hw/i2c/core: Add i2c_try_create_slave() and i2c_realize_and_unref() hw/misc/pca9552: Rename 'nr_leds' as 'pin_count' hw/misc/pca9552: Rename generic code as pca955x hw/misc/pca9552: Add generic PCA955xClass, parent of TYPE_PCA9552 hw/misc/pca9552: Add a 'description' property for debugging purpose hw/misc/pca9552: Trace GPIO High/Low events hw/arm/aspeed: Describe each PCA9552 device hw/misc/pca9552: Trace GPIO change events hw/misc/pca9552: Model qdev output GPIOs Richard Henderson (45): target/arm: Add isar tests for mte target/arm: Improve masking of SCR RES0 bits target/arm: Add support for MTE to SCTLR_ELx target/arm: Add support for MTE to HCR_EL2 and SCR_EL3 target/arm: Rename DISAS_UPDATE to DISAS_UPDATE_EXIT target/arm: Add DISAS_UPDATE_NOCHAIN target/arm: Add MTE system registers target/arm: Add MTE bits to tb_flags target/arm: Implement the IRG instruction target/arm: Revise decoding for disas_add_sub_imm target/arm: Implement the ADDG, SUBG instructions target/arm: Implement the GMI instruction target/arm: Implement the SUBP instruction target/arm: Define arm_cpu_do_unaligned_access for user-only target/arm: Implement LDG, STG, ST2G instructions target/arm: Implement the STGP instruction target/arm: Restrict the values of DCZID.BS under TCG target/arm: Simplify DC_ZVA target/arm: Implement the LDGM, STGM, STZGM instructions target/arm: Implement the access tag cache flushes target/arm: Move regime_el to internals.h target/arm: Move regime_tcr to internals.h target/arm: Add gen_mte_check1 target/arm: Add gen_mte_checkN target/arm: Implement helper_mte_check1 target/arm: Implement helper_mte_checkN target/arm: Add helper_mte_check_zva target/arm: Use mte_checkN for sve unpredicated loads target/arm: Use mte_checkN for sve unpredicated stores target/arm: Use mte_check1 for sve LD1R target/arm: Tidy trans_LD1R_zpri target/arm: Add arm_tlb_bti_gp target/arm: Add mte helpers for sve scalar + int loads target/arm: Add mte helpers for sve scalar + int stores target/arm: Add mte helpers for sve scalar + int ff/nf loads target/arm: Handle TBI for sve scalar + int memory ops target/arm: Add mte helpers for sve scatter/gather memory ops target/arm: Complete TBI clearing for user-only for SVE target/arm: Implement data cache set allocation tags target/arm: Set PSTATE.TCO on exception entry target/arm: Always pass cacheattr to get_phys_addr target/arm: Cache the Tagged bit for a page in MemTxAttrs target/arm: Create tagged ram when MTE is enabled target/arm: Add allocation tag storage for system mode target/arm: Enable MTE include/hw/arm/aspeed.h | 12 +- include/hw/i2c/i2c.h | 2 + include/hw/misc/pca9552.h | 16 +- target/arm/cpu.h | 50 +- target/arm/helper-a64.h | 16 + target/arm/helper-sve.h | 488 ++++++++++++++ target/arm/helper.h | 2 + target/arm/internals.h | 153 ++++- target/arm/translate-a64.h | 5 + target/arm/translate.h | 23 +- hw/arm/aspeed.c | 46 +- hw/arm/virt.c | 55 +- hw/i2c/core.c | 18 +- hw/misc/pca9552.c | 216 +++++-- target/arm/cpu.c | 81 ++- target/arm/cpu64.c | 5 + target/arm/helper-a64.c | 94 +-- target/arm/helper.c | 423 ++++++++++--- target/arm/m_helper.c | 11 +- target/arm/mte_helper.c | 906 ++++++++++++++++++++++++++ target/arm/op_helper.c | 16 + target/arm/sve_helper.c | 616 ++++++++++++++---- target/arm/tlb_helper.c | 13 +- target/arm/translate-a64.c | 657 ++++++++++++++++--- target/arm/translate-sve.c | 1366 ++++++++++++++++++++++++++-------------- target/arm/translate-vfp.inc.c | 4 +- target/arm/translate.c | 16 +- hw/misc/trace-events | 4 + target/arm/Makefile.objs | 1 + 29 files changed, 4391 insertions(+), 924 deletions(-) create mode 100644 target/arm/mte_helper.c From: Philippe Mathieu-Daudé <f4bug@amsat.org> I'm confused by this code, 'bmc' is created as: bmc = g_new0(AspeedBoardState, 1); Then we use it as QOM owner for different MemoryRegion objects. But looking at memory_region_init_ram (similarly for ROM): void memory_region_init_ram(MemoryRegion *mr, struct Object *owner, const char *name, uint64_t size, Error **errp) { DeviceState *owner_dev; Error *err = NULL; memory_region_init_ram_nomigrate(mr, owner, name, size, &err); if (err) { error_propagate(errp, err); return; } /* This will assert if owner is neither NULL nor a DeviceState. * We only want the owner here for the purposes of defining a * unique name for migration. TODO: Ideally we should implement * a naming scheme for Objects which are not DeviceStates, in * which case we can relax this restriction. */ owner_dev = DEVICE(owner); vmstate_register_ram(mr, owner_dev); } The expected assertion is not triggered ('bmc' is not NULL neither a DeviceState). 'bmc' structure is defined as: struct AspeedBoardState { AspeedSoCState soc; MemoryRegion ram_container; MemoryRegion max_ram; }; What happens is when using 'OBJECT(bmc)', the QOM macros cast the memory pointed by bmc, which first member is 'soc', which is initialized ...: object_initialize_child(OBJECT(machine), "soc", &bmc->soc, amc->soc_name); The 'soc' object is indeed a DeviceState, so the assertion passes. Since this is fragile and only happens to work by luck, remove the dangerous OBJECT(bmc) owner argument. Note, this probably breaks migration for this machine. Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200623072132.2868-2-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/aspeed.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/hw/arm/aspeed.c b/hw/arm/aspeed.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/aspeed.c +++ b/hw/arm/aspeed.c @@ -XXX,XX +XXX,XX @@ static void aspeed_machine_init(MachineState *machine) * needed by the flash modules of the Aspeed machines. */ if (ASPEED_MACHINE(machine)->mmio_exec) { - memory_region_init_alias(boot_rom, OBJECT(bmc), "aspeed.boot_rom", + memory_region_init_alias(boot_rom, NULL, "aspeed.boot_rom", &fl->mmio, 0, fl->size); memory_region_add_subregion(get_system_memory(), FIRMWARE_ADDR, boot_rom); } else { - memory_region_init_rom(boot_rom, OBJECT(bmc), "aspeed.boot_rom", + memory_region_init_rom(boot_rom, NULL, "aspeed.boot_rom", fl->size, &error_abort); memory_region_add_subregion(get_system_memory(), FIRMWARE_ADDR, boot_rom); @@ -XXX,XX +XXX,XX @@ static void aspeed_machine_init(MachineState *machine) if (machine->kernel_filename && sc->num_cpus > 1) { /* With no u-boot we must set up a boot stub for the secondary CPU */ MemoryRegion *smpboot = g_new(MemoryRegion, 1); - memory_region_init_ram(smpboot, OBJECT(bmc), "aspeed.smpboot", + memory_region_init_ram(smpboot, NULL, "aspeed.smpboot", 0x80, &error_abort); memory_region_add_subregion(get_system_memory(), AST_SMP_MAILBOX_BASE, smpboot); -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> To have a more consistent naming, rename AspeedBoardState as AspeedMachineState. Suggested-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072132.2868-3-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/aspeed.h | 4 ++-- hw/arm/aspeed.c | 20 ++++++++++---------- 2 files changed, 12 insertions(+), 12 deletions(-) diff --git a/include/hw/arm/aspeed.h b/include/hw/arm/aspeed.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/aspeed.h +++ b/include/hw/arm/aspeed.h @@ -XXX,XX +XXX,XX @@ #include "hw/boards.h" -typedef struct AspeedBoardState AspeedBoardState; +typedef struct AspeedMachineState AspeedMachineState; #define TYPE_ASPEED_MACHINE MACHINE_TYPE_NAME("aspeed") #define ASPEED_MACHINE(obj) \ @@ -XXX,XX +XXX,XX @@ typedef struct AspeedMachineClass { const char *spi_model; uint32_t num_cs; uint32_t macs_mask; - void (*i2c_init)(AspeedBoardState *bmc); + void (*i2c_init)(AspeedMachineState *bmc); } AspeedMachineClass; diff --git a/hw/arm/aspeed.c b/hw/arm/aspeed.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/aspeed.c +++ b/hw/arm/aspeed.c @@ -XXX,XX +XXX,XX @@ static struct arm_boot_info aspeed_board_binfo = { .board_id = -1, /* device-tree-only board */ }; -struct AspeedBoardState { +struct AspeedMachineState { AspeedSoCState soc; MemoryRegion ram_container; MemoryRegion max_ram; @@ -XXX,XX +XXX,XX @@ static void sdhci_attach_drive(SDHCIState *sdhci, DriveInfo *dinfo) static void aspeed_machine_init(MachineState *machine) { - AspeedBoardState *bmc; + AspeedMachineState *bmc; AspeedMachineClass *amc = ASPEED_MACHINE_GET_CLASS(machine); AspeedSoCClass *sc; DriveInfo *drive0 = drive_get(IF_MTD, 0, 0); @@ -XXX,XX +XXX,XX @@ static void aspeed_machine_init(MachineState *machine) int i; NICInfo *nd = &nd_table[0]; - bmc = g_new0(AspeedBoardState, 1); + bmc = g_new0(AspeedMachineState, 1); memory_region_init(&bmc->ram_container, NULL, "aspeed-ram-container", 4 * GiB); @@ -XXX,XX +XXX,XX @@ static void aspeed_machine_init(MachineState *machine) arm_load_kernel(ARM_CPU(first_cpu), machine, &aspeed_board_binfo); } -static void palmetto_bmc_i2c_init(AspeedBoardState *bmc) +static void palmetto_bmc_i2c_init(AspeedMachineState *bmc) { AspeedSoCState *soc = &bmc->soc; DeviceState *dev; @@ -XXX,XX +XXX,XX @@ static void palmetto_bmc_i2c_init(AspeedBoardState *bmc) object_property_set_int(OBJECT(dev), 110000, "temperature3", &error_abort); } -static void ast2500_evb_i2c_init(AspeedBoardState *bmc) +static void ast2500_evb_i2c_init(AspeedMachineState *bmc) { AspeedSoCState *soc = &bmc->soc; uint8_t *eeprom_buf = g_malloc0(8 * 1024); @@ -XXX,XX +XXX,XX @@ static void ast2500_evb_i2c_init(AspeedBoardState *bmc) i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 11), "ds1338", 0x32); } -static void ast2600_evb_i2c_init(AspeedBoardState *bmc) +static void ast2600_evb_i2c_init(AspeedMachineState *bmc) { /* Start with some devices on our I2C busses */ ast2500_evb_i2c_init(bmc); } -static void romulus_bmc_i2c_init(AspeedBoardState *bmc) +static void romulus_bmc_i2c_init(AspeedMachineState *bmc) { AspeedSoCState *soc = &bmc->soc; @@ -XXX,XX +XXX,XX @@ static void romulus_bmc_i2c_init(AspeedBoardState *bmc) i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 11), "ds1338", 0x32); } -static void swift_bmc_i2c_init(AspeedBoardState *bmc) +static void swift_bmc_i2c_init(AspeedMachineState *bmc) { AspeedSoCState *soc = &bmc->soc; @@ -XXX,XX +XXX,XX @@ static void swift_bmc_i2c_init(AspeedBoardState *bmc) i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 12), "tmp105", 0x4a); } -static void sonorapass_bmc_i2c_init(AspeedBoardState *bmc) +static void sonorapass_bmc_i2c_init(AspeedMachineState *bmc) { AspeedSoCState *soc = &bmc->soc; @@ -XXX,XX +XXX,XX @@ static void sonorapass_bmc_i2c_init(AspeedBoardState *bmc) } -static void witherspoon_bmc_i2c_init(AspeedBoardState *bmc) +static void witherspoon_bmc_i2c_init(AspeedMachineState *bmc) { AspeedSoCState *soc = &bmc->soc; uint8_t *eeprom_buf = g_malloc0(8 * 1024); -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> AspeedMachineState seems crippled. We use incorrectly 2 different structures to do the same thing. Merge them altogether: - Move AspeedMachine fields to AspeedMachineState - AspeedMachineState is now QOM - Remove unused AspeedMachine structure Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072132.2868-4-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/aspeed.h | 8 +------- hw/arm/aspeed.c | 11 +++++++---- 2 files changed, 8 insertions(+), 11 deletions(-) diff --git a/include/hw/arm/aspeed.h b/include/hw/arm/aspeed.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/aspeed.h +++ b/include/hw/arm/aspeed.h @@ -XXX,XX +XXX,XX @@ typedef struct AspeedMachineState AspeedMachineState; #define TYPE_ASPEED_MACHINE MACHINE_TYPE_NAME("aspeed") #define ASPEED_MACHINE(obj) \ - OBJECT_CHECK(AspeedMachine, (obj), TYPE_ASPEED_MACHINE) - -typedef struct AspeedMachine { - MachineState parent_obj; - - bool mmio_exec; -} AspeedMachine; + OBJECT_CHECK(AspeedMachineState, (obj), TYPE_ASPEED_MACHINE) #define ASPEED_MAC0_ON (1 << 0) #define ASPEED_MAC1_ON (1 << 1) diff --git a/hw/arm/aspeed.c b/hw/arm/aspeed.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/aspeed.c +++ b/hw/arm/aspeed.c @@ -XXX,XX +XXX,XX @@ static struct arm_boot_info aspeed_board_binfo = { }; struct AspeedMachineState { + /* Private */ + MachineState parent_obj; + /* Public */ + AspeedSoCState soc; MemoryRegion ram_container; MemoryRegion max_ram; + bool mmio_exec; }; /* Palmetto hardware value: 0x120CE416 */ @@ -XXX,XX +XXX,XX @@ static void sdhci_attach_drive(SDHCIState *sdhci, DriveInfo *dinfo) static void aspeed_machine_init(MachineState *machine) { - AspeedMachineState *bmc; + AspeedMachineState *bmc = ASPEED_MACHINE(machine); AspeedMachineClass *amc = ASPEED_MACHINE_GET_CLASS(machine); AspeedSoCClass *sc; DriveInfo *drive0 = drive_get(IF_MTD, 0, 0); @@ -XXX,XX +XXX,XX @@ static void aspeed_machine_init(MachineState *machine) int i; NICInfo *nd = &nd_table[0]; - bmc = g_new0(AspeedMachineState, 1); - memory_region_init(&bmc->ram_container, NULL, "aspeed-ram-container", 4 * GiB); memory_region_add_subregion(&bmc->ram_container, 0, machine->ram); @@ -XXX,XX +XXX,XX @@ static const TypeInfo aspeed_machine_types[] = { }, { .name = TYPE_ASPEED_MACHINE, .parent = TYPE_MACHINE, - .instance_size = sizeof(AspeedMachine), + .instance_size = sizeof(AspeedMachineState), .instance_init = aspeed_machine_instance_init, .class_size = sizeof(AspeedMachineClass), .class_init = aspeed_machine_class_init, -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Extract i2c_try_create_slave() and i2c_realize_and_unref() from i2c_create_slave(). We can now set properties on a I2CSlave before it is realized. This is in line with the recent qdev/QOM changes merged in commit 6675a653d2e. Reviewed-by: Corey Minyard <cminyard@mvista.com> Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Markus Armbruster <armbru@redhat.com> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-2-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/i2c/i2c.h | 2 ++ hw/i2c/core.c | 18 ++++++++++++++++-- 2 files changed, 18 insertions(+), 2 deletions(-) diff --git a/include/hw/i2c/i2c.h b/include/hw/i2c/i2c.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/i2c/i2c.h +++ b/include/hw/i2c/i2c.h @@ -XXX,XX +XXX,XX @@ int i2c_send(I2CBus *bus, uint8_t data); uint8_t i2c_recv(I2CBus *bus); DeviceState *i2c_create_slave(I2CBus *bus, const char *name, uint8_t addr); +DeviceState *i2c_try_create_slave(const char *name, uint8_t addr); +bool i2c_realize_and_unref(DeviceState *dev, I2CBus *bus, Error **errp); /* lm832x.c */ void lm832x_key_event(DeviceState *dev, int key, int state); diff --git a/hw/i2c/core.c b/hw/i2c/core.c index XXXXXXX..XXXXXXX 100644 --- a/hw/i2c/core.c +++ b/hw/i2c/core.c @@ -XXX,XX +XXX,XX @@ const VMStateDescription vmstate_i2c_slave = { } }; -DeviceState *i2c_create_slave(I2CBus *bus, const char *name, uint8_t addr) +DeviceState *i2c_try_create_slave(const char *name, uint8_t addr) { DeviceState *dev; dev = qdev_new(name); qdev_prop_set_uint8(dev, "address", addr); - qdev_realize_and_unref(dev, &bus->qbus, &error_fatal); + return dev; +} + +bool i2c_realize_and_unref(DeviceState *dev, I2CBus *bus, Error **errp) +{ + return qdev_realize_and_unref(dev, &bus->qbus, errp); +} + +DeviceState *i2c_create_slave(I2CBus *bus, const char *name, uint8_t addr) +{ + DeviceState *dev; + + dev = i2c_try_create_slave(name, addr); + i2c_realize_and_unref(dev, bus, &error_fatal); + return dev; } -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> The PCA9552 device does not expose LEDs, but simple pins to connnect LEDs to. To be clearer with the device model, rename 'nr_leds' as 'pin_count'. Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-3-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/misc/pca9552.h | 2 +- hw/misc/pca9552.c | 10 +++++----- 2 files changed, 6 insertions(+), 6 deletions(-) diff --git a/include/hw/misc/pca9552.h b/include/hw/misc/pca9552.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/pca9552.h +++ b/include/hw/misc/pca9552.h @@ -XXX,XX +XXX,XX @@ typedef struct PCA9552State { uint8_t regs[PCA9552_NR_REGS]; uint8_t max_reg; - uint8_t nr_leds; + uint8_t pin_count; } PCA9552State; #endif diff --git a/hw/misc/pca9552.c b/hw/misc/pca9552.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/pca9552.c +++ b/hw/misc/pca9552.c @@ -XXX,XX +XXX,XX @@ static void pca9552_update_pin_input(PCA9552State *s) { int i; - for (i = 0; i < s->nr_leds; i++) { + for (i = 0; i < s->pin_count; i++) { uint8_t input_reg = PCA9552_INPUT0 + (i / 8); uint8_t input_shift = (i % 8); uint8_t config = pca9552_pin_get_config(s, i); @@ -XXX,XX +XXX,XX @@ static void pca9552_get_led(Object *obj, Visitor *v, const char *name, error_setg(errp, "%s: error reading %s", __func__, name); return; } - if (led < 0 || led > s->nr_leds) { + if (led < 0 || led > s->pin_count) { error_setg(errp, "%s invalid led %s", __func__, name); return; } @@ -XXX,XX +XXX,XX @@ static void pca9552_set_led(Object *obj, Visitor *v, const char *name, error_setg(errp, "%s: error reading %s", __func__, name); return; } - if (led < 0 || led > s->nr_leds) { + if (led < 0 || led > s->pin_count) { error_setg(errp, "%s invalid led %s", __func__, name); return; } @@ -XXX,XX +XXX,XX @@ static void pca9552_initfn(Object *obj) * PCA955X device */ s->max_reg = PCA9552_LS3; - s->nr_leds = 16; + s->pin_count = 16; - for (led = 0; led < s->nr_leds; led++) { + for (led = 0; led < s->pin_count; led++) { char *name; name = g_strdup_printf("led%d", led); -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Various code from the PCA9552 device model is generic to the PCA955X family. We'll split the generic code in a base class in the next commit. To ease review, first do a dumb renaming. Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-4-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/misc/pca9552.h | 10 ++--- hw/misc/pca9552.c | 80 +++++++++++++++++++-------------------- 2 files changed, 45 insertions(+), 45 deletions(-) diff --git a/include/hw/misc/pca9552.h b/include/hw/misc/pca9552.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/pca9552.h +++ b/include/hw/misc/pca9552.h @@ -XXX,XX +XXX,XX @@ #include "hw/i2c/i2c.h" #define TYPE_PCA9552 "pca9552" -#define PCA9552(obj) OBJECT_CHECK(PCA9552State, (obj), TYPE_PCA9552) +#define PCA955X(obj) OBJECT_CHECK(PCA955xState, (obj), TYPE_PCA9552) -#define PCA9552_NR_REGS 10 +#define PCA955X_NR_REGS 10 -typedef struct PCA9552State { +typedef struct PCA955xState { /*< private >*/ I2CSlave i2c; /*< public >*/ @@ -XXX,XX +XXX,XX @@ typedef struct PCA9552State { uint8_t len; uint8_t pointer; - uint8_t regs[PCA9552_NR_REGS]; + uint8_t regs[PCA955X_NR_REGS]; uint8_t max_reg; uint8_t pin_count; -} PCA9552State; +} PCA955xState; #endif diff --git a/hw/misc/pca9552.c b/hw/misc/pca9552.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/pca9552.c +++ b/hw/misc/pca9552.c @@ -XXX,XX +XXX,XX @@ static const char *led_state[] = {"on", "off", "pwm0", "pwm1"}; -static uint8_t pca9552_pin_get_config(PCA9552State *s, int pin) +static uint8_t pca955x_pin_get_config(PCA955xState *s, int pin) { uint8_t reg = PCA9552_LS0 + (pin / 4); uint8_t shift = (pin % 4) << 1; @@ -XXX,XX +XXX,XX @@ static uint8_t pca9552_pin_get_config(PCA9552State *s, int pin) return extract32(s->regs[reg], shift, 2); } -static void pca9552_update_pin_input(PCA9552State *s) +static void pca955x_update_pin_input(PCA955xState *s) { int i; for (i = 0; i < s->pin_count; i++) { uint8_t input_reg = PCA9552_INPUT0 + (i / 8); uint8_t input_shift = (i % 8); - uint8_t config = pca9552_pin_get_config(s, i); + uint8_t config = pca955x_pin_get_config(s, i); switch (config) { case PCA9552_LED_ON: @@ -XXX,XX +XXX,XX @@ static void pca9552_update_pin_input(PCA9552State *s) } } -static uint8_t pca9552_read(PCA9552State *s, uint8_t reg) +static uint8_t pca955x_read(PCA955xState *s, uint8_t reg) { switch (reg) { case PCA9552_INPUT0: @@ -XXX,XX +XXX,XX @@ static uint8_t pca9552_read(PCA9552State *s, uint8_t reg) } } -static void pca9552_write(PCA9552State *s, uint8_t reg, uint8_t data) +static void pca955x_write(PCA955xState *s, uint8_t reg, uint8_t data) { switch (reg) { case PCA9552_PSC0: @@ -XXX,XX +XXX,XX @@ static void pca9552_write(PCA9552State *s, uint8_t reg, uint8_t data) case PCA9552_LS2: case PCA9552_LS3: s->regs[reg] = data; - pca9552_update_pin_input(s); + pca955x_update_pin_input(s); break; case PCA9552_INPUT0: @@ -XXX,XX +XXX,XX @@ static void pca9552_write(PCA9552State *s, uint8_t reg, uint8_t data) * after each byte is sent to or received by the device. The index * rollovers to 0 when the maximum register address is reached. */ -static void pca9552_autoinc(PCA9552State *s) +static void pca955x_autoinc(PCA955xState *s) { if (s->pointer != 0xFF && s->pointer & PCA9552_AUTOINC) { uint8_t reg = s->pointer & 0xf; @@ -XXX,XX +XXX,XX @@ static void pca9552_autoinc(PCA9552State *s) } } -static uint8_t pca9552_recv(I2CSlave *i2c) +static uint8_t pca955x_recv(I2CSlave *i2c) { - PCA9552State *s = PCA9552(i2c); + PCA955xState *s = PCA955X(i2c); uint8_t ret; - ret = pca9552_read(s, s->pointer & 0xf); + ret = pca955x_read(s, s->pointer & 0xf); /* * From the Specs: @@ -XXX,XX +XXX,XX @@ static uint8_t pca9552_recv(I2CSlave *i2c) __func__); } - pca9552_autoinc(s); + pca955x_autoinc(s); return ret; } -static int pca9552_send(I2CSlave *i2c, uint8_t data) +static int pca955x_send(I2CSlave *i2c, uint8_t data) { - PCA9552State *s = PCA9552(i2c); + PCA955xState *s = PCA955X(i2c); /* First byte sent by is the register address */ if (s->len == 0) { s->pointer = data; s->len++; } else { - pca9552_write(s, s->pointer & 0xf, data); + pca955x_write(s, s->pointer & 0xf, data); - pca9552_autoinc(s); + pca955x_autoinc(s); } return 0; } -static int pca9552_event(I2CSlave *i2c, enum i2c_event event) +static int pca955x_event(I2CSlave *i2c, enum i2c_event event) { - PCA9552State *s = PCA9552(i2c); + PCA955xState *s = PCA955X(i2c); s->len = 0; return 0; } -static void pca9552_get_led(Object *obj, Visitor *v, const char *name, +static void pca955x_get_led(Object *obj, Visitor *v, const char *name, void *opaque, Error **errp) { - PCA9552State *s = PCA9552(obj); + PCA955xState *s = PCA955X(obj); int led, rc, reg; uint8_t state; @@ -XXX,XX +XXX,XX @@ static void pca9552_get_led(Object *obj, Visitor *v, const char *name, * reading the INPUTx reg */ reg = PCA9552_LS0 + led / 4; - state = (pca9552_read(s, reg) >> (led % 8)) & 0x3; + state = (pca955x_read(s, reg) >> (led % 8)) & 0x3; visit_type_str(v, name, (char **)&led_state[state], errp); } @@ -XXX,XX +XXX,XX @@ static inline uint8_t pca955x_ledsel(uint8_t oldval, int led_num, int state) ((state & 0x3) << (led_num << 1)); } -static void pca9552_set_led(Object *obj, Visitor *v, const char *name, +static void pca955x_set_led(Object *obj, Visitor *v, const char *name, void *opaque, Error **errp) { - PCA9552State *s = PCA9552(obj); + PCA955xState *s = PCA955X(obj); Error *local_err = NULL; int led, rc, reg, val; uint8_t state; @@ -XXX,XX +XXX,XX @@ static void pca9552_set_led(Object *obj, Visitor *v, const char *name, } reg = PCA9552_LS0 + led / 4; - val = pca9552_read(s, reg); + val = pca955x_read(s, reg); val = pca955x_ledsel(val, led % 4, state); - pca9552_write(s, reg, val); + pca955x_write(s, reg, val); } static const VMStateDescription pca9552_vmstate = { @@ -XXX,XX +XXX,XX @@ static const VMStateDescription pca9552_vmstate = { .version_id = 0, .minimum_version_id = 0, .fields = (VMStateField[]) { - VMSTATE_UINT8(len, PCA9552State), - VMSTATE_UINT8(pointer, PCA9552State), - VMSTATE_UINT8_ARRAY(regs, PCA9552State, PCA9552_NR_REGS), - VMSTATE_I2C_SLAVE(i2c, PCA9552State), + VMSTATE_UINT8(len, PCA955xState), + VMSTATE_UINT8(pointer, PCA955xState), + VMSTATE_UINT8_ARRAY(regs, PCA955xState, PCA955X_NR_REGS), + VMSTATE_I2C_SLAVE(i2c, PCA955xState), VMSTATE_END_OF_LIST() } }; static void pca9552_reset(DeviceState *dev) { - PCA9552State *s = PCA9552(dev); + PCA955xState *s = PCA955X(dev); s->regs[PCA9552_PSC0] = 0xFF; s->regs[PCA9552_PWM0] = 0x80; @@ -XXX,XX +XXX,XX @@ static void pca9552_reset(DeviceState *dev) s->regs[PCA9552_LS2] = 0x55; s->regs[PCA9552_LS3] = 0x55; - pca9552_update_pin_input(s); + pca955x_update_pin_input(s); s->pointer = 0xFF; s->len = 0; } -static void pca9552_initfn(Object *obj) +static void pca955x_initfn(Object *obj) { - PCA9552State *s = PCA9552(obj); + PCA955xState *s = PCA955X(obj); int led; /* If support for the other PCA955X devices are implemented, these @@ -XXX,XX +XXX,XX @@ static void pca9552_initfn(Object *obj) char *name; name = g_strdup_printf("led%d", led); - object_property_add(obj, name, "bool", pca9552_get_led, pca9552_set_led, + object_property_add(obj, name, "bool", pca955x_get_led, pca955x_set_led, NULL, NULL); g_free(name); } @@ -XXX,XX +XXX,XX @@ static void pca9552_class_init(ObjectClass *klass, void *data) DeviceClass *dc = DEVICE_CLASS(klass); I2CSlaveClass *k = I2C_SLAVE_CLASS(klass); - k->event = pca9552_event; - k->recv = pca9552_recv; - k->send = pca9552_send; + k->event = pca955x_event; + k->recv = pca955x_recv; + k->send = pca955x_send; dc->reset = pca9552_reset; dc->vmsd = &pca9552_vmstate; } @@ -XXX,XX +XXX,XX @@ static void pca9552_class_init(ObjectClass *klass, void *data) static const TypeInfo pca9552_info = { .name = TYPE_PCA9552, .parent = TYPE_I2C_SLAVE, - .instance_init = pca9552_initfn, - .instance_size = sizeof(PCA9552State), + .instance_init = pca955x_initfn, + .instance_size = sizeof(PCA955xState), .class_init = pca9552_class_init, }; -static void pca9552_register_types(void) +static void pca955x_register_types(void) { type_register_static(&pca9552_info); } -type_init(pca9552_register_types) +type_init(pca955x_register_types) -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Extract the code common to the PCA955x family in PCA955xClass, keeping the PCA9552 specific parts into pca9552_class_init(). Remove the 'TODO' comment added in commit 5141d4158cf. Suggested-by: Cédric Le Goater <clg@kaod.org> Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-5-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/misc/pca9552.h | 6 ++-- hw/misc/pca9552.c | 66 ++++++++++++++++++++++++++++----------- 2 files changed, 51 insertions(+), 21 deletions(-) diff --git a/include/hw/misc/pca9552.h b/include/hw/misc/pca9552.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/pca9552.h +++ b/include/hw/misc/pca9552.h @@ -XXX,XX +XXX,XX @@ #include "hw/i2c/i2c.h" #define TYPE_PCA9552 "pca9552" -#define PCA955X(obj) OBJECT_CHECK(PCA955xState, (obj), TYPE_PCA9552) +#define TYPE_PCA955X "pca955x" +#define PCA955X(obj) OBJECT_CHECK(PCA955xState, (obj), TYPE_PCA955X) #define PCA955X_NR_REGS 10 +#define PCA955X_PIN_COUNT_MAX 16 typedef struct PCA955xState { /*< private >*/ @@ -XXX,XX +XXX,XX @@ typedef struct PCA955xState { uint8_t pointer; uint8_t regs[PCA955X_NR_REGS]; - uint8_t max_reg; - uint8_t pin_count; } PCA955xState; #endif diff --git a/hw/misc/pca9552.c b/hw/misc/pca9552.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/pca9552.c +++ b/hw/misc/pca9552.c @@ -XXX,XX +XXX,XX @@ * https://www.nxp.com/docs/en/application-note/AN264.pdf * * Copyright (c) 2017-2018, IBM Corporation. + * Copyright (c) 2020 Philippe Mathieu-Daudé * * This work is licensed under the terms of the GNU GPL, version 2 or * later. See the COPYING file in the top-level directory. @@ -XXX,XX +XXX,XX @@ #include "qapi/error.h" #include "qapi/visitor.h" +typedef struct PCA955xClass { + /*< private >*/ + I2CSlaveClass parent_class; + /*< public >*/ + + uint8_t pin_count; + uint8_t max_reg; +} PCA955xClass; + +#define PCA955X_CLASS(klass) \ + OBJECT_CLASS_CHECK(PCA955xClass, (klass), TYPE_PCA955X) +#define PCA955X_GET_CLASS(obj) \ + OBJECT_GET_CLASS(PCA955xClass, (obj), TYPE_PCA955X) + #define PCA9552_LED_ON 0x0 #define PCA9552_LED_OFF 0x1 #define PCA9552_LED_PWM0 0x2 @@ -XXX,XX +XXX,XX @@ static uint8_t pca955x_pin_get_config(PCA955xState *s, int pin) static void pca955x_update_pin_input(PCA955xState *s) { + PCA955xClass *k = PCA955X_GET_CLASS(s); int i; - for (i = 0; i < s->pin_count; i++) { + for (i = 0; i < k->pin_count; i++) { uint8_t input_reg = PCA9552_INPUT0 + (i / 8); uint8_t input_shift = (i % 8); uint8_t config = pca955x_pin_get_config(s, i); @@ -XXX,XX +XXX,XX @@ static void pca955x_write(PCA955xState *s, uint8_t reg, uint8_t data) */ static void pca955x_autoinc(PCA955xState *s) { + PCA955xClass *k = PCA955X_GET_CLASS(s); + if (s->pointer != 0xFF && s->pointer & PCA9552_AUTOINC) { uint8_t reg = s->pointer & 0xf; - reg = (reg + 1) % (s->max_reg + 1); + reg = (reg + 1) % (k->max_reg + 1); s->pointer = reg | PCA9552_AUTOINC; } } @@ -XXX,XX +XXX,XX @@ static int pca955x_event(I2CSlave *i2c, enum i2c_event event) static void pca955x_get_led(Object *obj, Visitor *v, const char *name, void *opaque, Error **errp) { + PCA955xClass *k = PCA955X_GET_CLASS(obj); PCA955xState *s = PCA955X(obj); int led, rc, reg; uint8_t state; @@ -XXX,XX +XXX,XX @@ static void pca955x_get_led(Object *obj, Visitor *v, const char *name, error_setg(errp, "%s: error reading %s", __func__, name); return; } - if (led < 0 || led > s->pin_count) { + if (led < 0 || led > k->pin_count) { error_setg(errp, "%s invalid led %s", __func__, name); return; } @@ -XXX,XX +XXX,XX @@ static inline uint8_t pca955x_ledsel(uint8_t oldval, int led_num, int state) static void pca955x_set_led(Object *obj, Visitor *v, const char *name, void *opaque, Error **errp) { + PCA955xClass *k = PCA955X_GET_CLASS(obj); PCA955xState *s = PCA955X(obj); Error *local_err = NULL; int led, rc, reg, val; @@ -XXX,XX +XXX,XX @@ static void pca955x_set_led(Object *obj, Visitor *v, const char *name, error_setg(errp, "%s: error reading %s", __func__, name); return; } - if (led < 0 || led > s->pin_count) { + if (led < 0 || led > k->pin_count) { error_setg(errp, "%s invalid led %s", __func__, name); return; } @@ -XXX,XX +XXX,XX @@ static void pca9552_reset(DeviceState *dev) static void pca955x_initfn(Object *obj) { - PCA955xState *s = PCA955X(obj); + PCA955xClass *k = PCA955X_GET_CLASS(obj); int led; - /* If support for the other PCA955X devices are implemented, these - * constant values might be part of class structure describing the - * PCA955X device - */ - s->max_reg = PCA9552_LS3; - s->pin_count = 16; - - for (led = 0; led < s->pin_count; led++) { + assert(k->pin_count <= PCA955X_PIN_COUNT_MAX); + for (led = 0; led < k->pin_count; led++) { char *name; name = g_strdup_printf("led%d", led); @@ -XXX,XX +XXX,XX @@ static void pca955x_initfn(Object *obj) } } -static void pca9552_class_init(ObjectClass *klass, void *data) +static void pca955x_class_init(ObjectClass *klass, void *data) { - DeviceClass *dc = DEVICE_CLASS(klass); I2CSlaveClass *k = I2C_SLAVE_CLASS(klass); k->event = pca955x_event; k->recv = pca955x_recv; k->send = pca955x_send; +} + +static const TypeInfo pca955x_info = { + .name = TYPE_PCA955X, + .parent = TYPE_I2C_SLAVE, + .instance_init = pca955x_initfn, + .instance_size = sizeof(PCA955xState), + .class_init = pca955x_class_init, + .abstract = true, +}; + +static void pca9552_class_init(ObjectClass *oc, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(oc); + PCA955xClass *pc = PCA955X_CLASS(oc); + dc->reset = pca9552_reset; dc->vmsd = &pca9552_vmstate; + pc->max_reg = PCA9552_LS3; + pc->pin_count = 16; } static const TypeInfo pca9552_info = { .name = TYPE_PCA9552, - .parent = TYPE_I2C_SLAVE, - .instance_init = pca955x_initfn, - .instance_size = sizeof(PCA955xState), + .parent = TYPE_PCA955X, .class_init = pca9552_class_init, }; static void pca955x_register_types(void) { + type_register_static(&pca955x_info); type_register_static(&pca9552_info); } -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Add a description field to distinguish between multiple devices. Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-6-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/misc/pca9552.h | 1 + hw/misc/pca9552.c | 18 ++++++++++++++++++ 2 files changed, 19 insertions(+) diff --git a/include/hw/misc/pca9552.h b/include/hw/misc/pca9552.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/pca9552.h +++ b/include/hw/misc/pca9552.h @@ -XXX,XX +XXX,XX @@ typedef struct PCA955xState { uint8_t pointer; uint8_t regs[PCA955X_NR_REGS]; + char *description; /* For debugging purpose only */ } PCA955xState; #endif diff --git a/hw/misc/pca9552.c b/hw/misc/pca9552.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/pca9552.c +++ b/hw/misc/pca9552.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "qemu/log.h" #include "qemu/module.h" +#include "hw/qdev-properties.h" #include "hw/misc/pca9552.h" #include "hw/misc/pca9552_regs.h" #include "migration/vmstate.h" @@ -XXX,XX +XXX,XX @@ static void pca955x_initfn(Object *obj) } } +static void pca955x_realize(DeviceState *dev, Error **errp) +{ + PCA955xState *s = PCA955X(dev); + + if (!s->description) { + s->description = g_strdup("pca-unspecified"); + } +} + +static Property pca955x_properties[] = { + DEFINE_PROP_STRING("description", PCA955xState, description), + DEFINE_PROP_END_OF_LIST(), +}; + static void pca955x_class_init(ObjectClass *klass, void *data) { + DeviceClass *dc = DEVICE_CLASS(klass); I2CSlaveClass *k = I2C_SLAVE_CLASS(klass); k->event = pca955x_event; k->recv = pca955x_recv; k->send = pca955x_send; + dc->realize = pca955x_realize; + device_class_set_props(dc, pca955x_properties); } static const TypeInfo pca955x_info = { -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Add a trivial representation of the PCA9552 GPIOs. Example booting obmc-phosphor-image: $ qemu-system-arm -M witherspoon-bmc -trace pca955x_gpio_status 1592689902.327837:pca955x_gpio_status pca-unspecified GPIOs 0-15 [*...............] 1592689902.329934:pca955x_gpio_status pca-unspecified GPIOs 0-15 [**..............] 1592689902.330717:pca955x_gpio_status pca-unspecified GPIOs 0-15 [***.............] 1592689902.331431:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****............] 1592689902.332163:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........*..] 1592689902.332888:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........**.] 1592689902.333629:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........***] 1592690032.793289:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........*.*] 1592690033.303163:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........***] 1592690033.812962:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........*.*] 1592690034.323234:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........***] 1592690034.832922:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........*.*] We notice the GPIO #14 (front-power LED) starts to blink. This LED is described in the witherspoon device-tree [*]: front-power { retain-state-shutdown; default-state = "keep"; gpios = <&pca0 14 GPIO_ACTIVE_LOW>; }; [*] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/arm/boot/dts/aspeed-bmc-opp-witherspoon.dts?id=b1f9be9392f0#n140 Suggested-by: Cédric Le Goater <clg@kaod.org> Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-7-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/misc/pca9552.c | 39 +++++++++++++++++++++++++++++++++++++++ hw/misc/trace-events | 3 +++ 2 files changed, 42 insertions(+) diff --git a/hw/misc/pca9552.c b/hw/misc/pca9552.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/pca9552.c +++ b/hw/misc/pca9552.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "qemu/log.h" #include "qemu/module.h" +#include "qemu/bitops.h" #include "hw/qdev-properties.h" #include "hw/misc/pca9552.h" #include "hw/misc/pca9552_regs.h" #include "migration/vmstate.h" #include "qapi/error.h" #include "qapi/visitor.h" +#include "trace.h" typedef struct PCA955xClass { /*< private >*/ @@ -XXX,XX +XXX,XX @@ static uint8_t pca955x_pin_get_config(PCA955xState *s, int pin) return extract32(s->regs[reg], shift, 2); } +/* Return INPUT status (bit #N belongs to GPIO #N) */ +static uint16_t pca955x_pins_get_status(PCA955xState *s) +{ + return (s->regs[PCA9552_INPUT1] << 8) | s->regs[PCA9552_INPUT0]; +} + +static void pca955x_display_pins_status(PCA955xState *s, + uint16_t previous_pins_status) +{ + PCA955xClass *k = PCA955X_GET_CLASS(s); + uint16_t pins_status, pins_changed; + int i; + + pins_status = pca955x_pins_get_status(s); + pins_changed = previous_pins_status ^ pins_status; + if (!pins_changed) { + return; + } + if (trace_event_get_state_backends(TRACE_PCA955X_GPIO_STATUS)) { + char *buf = g_newa(char, k->pin_count + 1); + + for (i = 0; i < k->pin_count; i++) { + if (extract32(pins_status, i, 1)) { + buf[i] = '*'; + } else { + buf[i] = '.'; + } + } + buf[i] = '\0'; + trace_pca955x_gpio_status(s->description, buf); + } +} + static void pca955x_update_pin_input(PCA955xState *s) { PCA955xClass *k = PCA955X_GET_CLASS(s); @@ -XXX,XX +XXX,XX @@ static uint8_t pca955x_read(PCA955xState *s, uint8_t reg) static void pca955x_write(PCA955xState *s, uint8_t reg, uint8_t data) { + uint16_t pins_status; + switch (reg) { case PCA9552_PSC0: case PCA9552_PWM0: @@ -XXX,XX +XXX,XX @@ static void pca955x_write(PCA955xState *s, uint8_t reg, uint8_t data) case PCA9552_LS1: case PCA9552_LS2: case PCA9552_LS3: + pins_status = pca955x_pins_get_status(s); s->regs[reg] = data; pca955x_update_pin_input(s); + pca955x_display_pins_status(s, pins_status); break; case PCA9552_INPUT0: 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 @@ via1_adb_poll(uint8_t data, const char *vadbint, int status, int index, int size # grlib_ahb_apb_pnp.c grlib_ahb_pnp_read(uint64_t addr, uint32_t value) "AHB PnP read addr:0x%03"PRIx64" data:0x%08x" grlib_apb_pnp_read(uint64_t addr, uint32_t value) "APB PnP read addr:0x%03"PRIx64" data:0x%08x" + +# pca9552.c +pca955x_gpio_status(const char *description, const char *buf) "%s GPIOs 0-15 [%s]" -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> We have 2 distinct PCA9552 devices. Set their description to distinguish them when looking at the trace events. Description name taken from: https://github.com/open-power/witherspoon-xml/blob/master/witherspoon.xml Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Corey Minyard <cminyard@mvista.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-8-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/aspeed.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/hw/arm/aspeed.c b/hw/arm/aspeed.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/aspeed.c +++ b/hw/arm/aspeed.c @@ -XXX,XX +XXX,XX @@ static void witherspoon_bmc_i2c_init(AspeedMachineState *bmc) { AspeedSoCState *soc = &bmc->soc; uint8_t *eeprom_buf = g_malloc0(8 * 1024); + DeviceState *dev; /* Bus 3: TODO bmp280@77 */ /* Bus 3: TODO max31785@52 */ /* Bus 3: TODO dps310@76 */ - i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 3), TYPE_PCA9552, - 0x60); + dev = i2c_try_create_slave(TYPE_PCA9552, 0x60); + qdev_prop_set_string(dev, "description", "pca1"); + i2c_realize_and_unref(dev, aspeed_i2c_get_bus(DEVICE(&soc->i2c), 3), + &error_fatal); i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 4), "tmp423", 0x4c); i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 5), "tmp423", 0x4c); @@ -XXX,XX +XXX,XX @@ static void witherspoon_bmc_i2c_init(AspeedMachineState *bmc) smbus_eeprom_init_one(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 11), 0x51, eeprom_buf); - i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 11), TYPE_PCA9552, - 0x60); + dev = i2c_try_create_slave(TYPE_PCA9552, 0x60); + qdev_prop_set_string(dev, "description", "pca0"); + i2c_realize_and_unref(dev, aspeed_i2c_get_bus(DEVICE(&soc->i2c), 11), + &error_fatal); /* Bus 11: TODO ucd90160@64 */ } -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Emit a trace event when a GPIO change its state. Example booting obmc-phosphor-image: $ qemu-system-arm -M witherspoon-bmc -trace pca955x_gpio_change 1592690552.687372:pca955x_gpio_change pca1 GPIO id:0 status: 0 -> 1 1592690552.690169:pca955x_gpio_change pca1 GPIO id:1 status: 0 -> 1 1592690552.691673:pca955x_gpio_change pca1 GPIO id:2 status: 0 -> 1 1592690552.696886:pca955x_gpio_change pca1 GPIO id:3 status: 0 -> 1 1592690552.698614:pca955x_gpio_change pca1 GPIO id:13 status: 0 -> 1 1592690552.699833:pca955x_gpio_change pca1 GPIO id:14 status: 0 -> 1 1592690552.700842:pca955x_gpio_change pca1 GPIO id:15 status: 0 -> 1 1592690683.841921:pca955x_gpio_change pca1 GPIO id:14 status: 1 -> 0 1592690683.861660:pca955x_gpio_change pca1 GPIO id:14 status: 0 -> 1 1592690684.371460:pca955x_gpio_change pca1 GPIO id:14 status: 1 -> 0 1592690684.882115:pca955x_gpio_change pca1 GPIO id:14 status: 0 -> 1 1592690685.391411:pca955x_gpio_change pca1 GPIO id:14 status: 1 -> 0 1592690685.901391:pca955x_gpio_change pca1 GPIO id:14 status: 0 -> 1 1592690686.411678:pca955x_gpio_change pca1 GPIO id:14 status: 1 -> 0 1592690686.921279:pca955x_gpio_change pca1 GPIO id:14 status: 0 -> 1 We notice the GPIO #14 (front-power LED) starts to blink. This LED is described in the witherspoon device-tree [*]: front-power { retain-state-shutdown; default-state = "keep"; gpios = <&pca0 14 GPIO_ACTIVE_LOW>; }; [*] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/arm/boot/dts/aspeed-bmc-opp-witherspoon.dts?id=b1f9be9392f0#n140 Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-9-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/misc/pca9552.c | 15 +++++++++++++++ hw/misc/trace-events | 1 + 2 files changed, 16 insertions(+) diff --git a/hw/misc/pca9552.c b/hw/misc/pca9552.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/pca9552.c +++ b/hw/misc/pca9552.c @@ -XXX,XX +XXX,XX @@ static void pca955x_display_pins_status(PCA955xState *s, buf[i] = '\0'; trace_pca955x_gpio_status(s->description, buf); } + if (trace_event_get_state_backends(TRACE_PCA955X_GPIO_CHANGE)) { + for (i = 0; i < k->pin_count; i++) { + if (extract32(pins_changed, i, 1)) { + unsigned new_state = extract32(pins_status, i, 1); + + /* + * We display the state using the PCA logic ("active-high"). + * This is not the state of the LED, which signal might be + * wired "active-low" on the board. + */ + trace_pca955x_gpio_change(s->description, i, + !new_state, new_state); + } + } + } } static void pca955x_update_pin_input(PCA955xState *s) 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 @@ grlib_apb_pnp_read(uint64_t addr, uint32_t value) "APB PnP read addr:0x%03"PRIx6 # pca9552.c pca955x_gpio_status(const char *description, const char *buf) "%s GPIOs 0-15 [%s]" +pca955x_gpio_change(const char *description, unsigned id, unsigned prev_state, unsigned current_state) "%s GPIO id:%u status: %u -> %u" -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> The PCA9552 has 16 GPIOs which can be used as input, output or PWM mode. QEMU models the output GPIO with the qemu_irq type. Let the device expose the 16 GPIOs to allow us to later connect LEDs to these outputs. Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-10-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/misc/pca9552.h | 1 + hw/misc/pca9552.c | 6 ++++++ 2 files changed, 7 insertions(+) diff --git a/include/hw/misc/pca9552.h b/include/hw/misc/pca9552.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/pca9552.h +++ b/include/hw/misc/pca9552.h @@ -XXX,XX +XXX,XX @@ typedef struct PCA955xState { uint8_t pointer; uint8_t regs[PCA955X_NR_REGS]; + qemu_irq gpio[PCA955X_PIN_COUNT_MAX]; char *description; /* For debugging purpose only */ } PCA955xState; diff --git a/hw/misc/pca9552.c b/hw/misc/pca9552.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/pca9552.c +++ b/hw/misc/pca9552.c @@ -XXX,XX +XXX,XX @@ #include "hw/qdev-properties.h" #include "hw/misc/pca9552.h" #include "hw/misc/pca9552_regs.h" +#include "hw/irq.h" #include "migration/vmstate.h" #include "qapi/error.h" #include "qapi/visitor.h" @@ -XXX,XX +XXX,XX @@ static void pca955x_update_pin_input(PCA955xState *s) switch (config) { case PCA9552_LED_ON: + qemu_set_irq(s->gpio[i], 1); s->regs[input_reg] |= 1 << input_shift; break; case PCA9552_LED_OFF: + qemu_set_irq(s->gpio[i], 0); s->regs[input_reg] &= ~(1 << input_shift); break; case PCA9552_LED_PWM0: @@ -XXX,XX +XXX,XX @@ static void pca955x_initfn(Object *obj) static void pca955x_realize(DeviceState *dev, Error **errp) { + PCA955xClass *k = PCA955X_GET_CLASS(dev); PCA955xState *s = PCA955X(dev); if (!s->description) { s->description = g_strdup("pca-unspecified"); } + + qdev_init_gpio_out(dev, s->gpio, k->pin_count); } static Property pca955x_properties[] = { -- 2.20.1 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: 20200626033144.790098-2-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ static inline bool isar_feature_aa64_bti(const ARMISARegisters *id) return FIELD_EX64(id->id_aa64pfr1, ID_AA64PFR1, BT) != 0; } +static inline bool isar_feature_aa64_mte_insn_reg(const ARMISARegisters *id) +{ + return FIELD_EX64(id->id_aa64pfr1, ID_AA64PFR1, MTE) != 0; +} + +static inline bool isar_feature_aa64_mte(const ARMISARegisters *id) +{ + return FIELD_EX64(id->id_aa64pfr1, ID_AA64PFR1, MTE) >= 2; +} + static inline bool isar_feature_aa64_pmu_8_1(const ARMISARegisters *id) { return FIELD_EX64(id->id_aa64dfr0, ID_AA64DFR0, PMUVER) >= 4 && -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Protect reads of aa64 id registers with ARM_CP_STATE_AA64. Use this as a simpler test than arm_el_is_aa64, since EL3 cannot change mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-3-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void scr_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value) uint32_t valid_mask = 0x3fff; ARMCPU *cpu = env_archcpu(env); - if (arm_el_is_aa64(env, 3)) { + if (ri->state == ARM_CP_STATE_AA64) { value |= SCR_FW | SCR_AW; /* these two bits are RES1. */ valid_mask &= ~SCR_NET; + + if (cpu_isar_feature(aa64_lor, cpu)) { + valid_mask |= SCR_TLOR; + } + if (cpu_isar_feature(aa64_pauth, cpu)) { + valid_mask |= SCR_API | SCR_APK; + } } else { valid_mask &= ~(SCR_RW | SCR_ST); } @@ -XXX,XX +XXX,XX @@ static void scr_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value) valid_mask &= ~SCR_SMD; } } - if (cpu_isar_feature(aa64_lor, cpu)) { - valid_mask |= SCR_TLOR; - } - if (cpu_isar_feature(aa64_pauth, cpu)) { - valid_mask |= SCR_API | SCR_APK; - } /* Clear all-context RES0 bits. */ value &= valid_mask; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> This does not attempt to rectify all of the res0 bits, but does clear the mte bits when not enabled. Since there is no high-part mapping of SCTLR, aa32 mode cannot write to these bits. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-4-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 23 +++++++++++++++++------ 1 file changed, 17 insertions(+), 6 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void sctlr_write(CPUARMState *env, const ARMCPRegInfo *ri, { ARMCPU *cpu = env_archcpu(env); + if (arm_feature(env, ARM_FEATURE_PMSA) && !cpu->has_mpu) { + /* M bit is RAZ/WI for PMSA with no MPU implemented */ + value &= ~SCTLR_M; + } + + /* ??? Lots of these bits are not implemented. */ + + if (ri->state == ARM_CP_STATE_AA64 && !cpu_isar_feature(aa64_mte, cpu)) { + if (ri->opc1 == 6) { /* SCTLR_EL3 */ + value &= ~(SCTLR_ITFSB | SCTLR_TCF | SCTLR_ATA); + } else { + value &= ~(SCTLR_ITFSB | SCTLR_TCF0 | SCTLR_TCF | + SCTLR_ATA0 | SCTLR_ATA); + } + } + if (raw_read(env, ri) == value) { /* Skip the TLB flush if nothing actually changed; Linux likes * to do a lot of pointless SCTLR writes. @@ -XXX,XX +XXX,XX @@ static void sctlr_write(CPUARMState *env, const ARMCPRegInfo *ri, return; } - if (arm_feature(env, ARM_FEATURE_PMSA) && !cpu->has_mpu) { - /* M bit is RAZ/WI for PMSA with no MPU implemented */ - value &= ~SCTLR_M; - } - raw_write(env, ri, value); - /* ??? Lots of these bits are not implemented. */ + /* This may enable/disable the MMU, so do a TLB flush. */ tlb_flush(CPU(cpu)); -- 2.20.1 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: 20200626033144.790098-5-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void scr_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value) if (cpu_isar_feature(aa64_pauth, cpu)) { valid_mask |= SCR_API | SCR_APK; } + if (cpu_isar_feature(aa64_mte, cpu)) { + valid_mask |= SCR_ATA; + } } else { valid_mask &= ~(SCR_RW | SCR_ST); } @@ -XXX,XX +XXX,XX @@ static void do_hcr_write(CPUARMState *env, uint64_t value, uint64_t valid_mask) if (cpu_isar_feature(aa64_pauth, cpu)) { valid_mask |= HCR_API | HCR_APK; } + if (cpu_isar_feature(aa64_mte, cpu)) { + valid_mask |= HCR_ATA | HCR_DCT | HCR_TID5; + } } /* Clear RES0 bits. */ value &= valid_mask; - /* These bits change the MMU setup: + /* + * These bits change the MMU setup: * HCR_VM enables stage 2 translation * HCR_PTW forbids certain page-table setups - * HCR_DC Disables stage1 and enables stage2 translation + * HCR_DC disables stage1 and enables stage2 translation + * HCR_DCT enables tagging on (disabled) stage1 translation */ - if ((env->cp15.hcr_el2 ^ value) & (HCR_VM | HCR_PTW | HCR_DC)) { + if ((env->cp15.hcr_el2 ^ value) & (HCR_VM | HCR_PTW | HCR_DC | HCR_DCT)) { tlb_flush(CPU(cpu)); } env->cp15.hcr_el2 = value; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Emphasize that the is_jmp option exits to the main loop. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-6-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.h | 14 ++++++++------ target/arm/translate-a64.c | 8 ++++---- target/arm/translate-vfp.inc.c | 4 ++-- target/arm/translate.c | 12 ++++++------ 4 files changed, 20 insertions(+), 18 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 @@ static inline void disas_set_insn_syndrome(DisasContext *s, uint32_t syn) /* is_jmp field values */ #define DISAS_JUMP DISAS_TARGET_0 /* only pc was modified dynamically */ -#define DISAS_UPDATE DISAS_TARGET_1 /* cpu state was modified dynamically */ +/* CPU state was modified dynamically; exit to main loop for interrupts. */ +#define DISAS_UPDATE_EXIT DISAS_TARGET_1 /* These instructions trap after executing, so the A32/T32 decoder must * defer them until after the conditional execution state has been updated. * WFI also needs special handling when single-stepping. @@ -XXX,XX +XXX,XX @@ static inline void disas_set_insn_syndrome(DisasContext *s, uint32_t syn) * custom end-of-TB code) */ #define DISAS_BX_EXCRET DISAS_TARGET_8 -/* For instructions which want an immediate exit to the main loop, - * as opposed to attempting to use lookup_and_goto_ptr. Unlike - * DISAS_UPDATE this doesn't write the PC on exiting the translation - * loop so you need to ensure something (gen_a64_set_pc_im or runtime - * helper) has done so before we reach return from cpu_tb_exec. +/* + * For instructions which want an immediate exit to the main loop, as opposed + * to attempting to use lookup_and_goto_ptr. Unlike DISAS_UPDATE_EXIT, this + * doesn't write the PC on exiting the translation loop so you need to ensure + * something (gen_a64_set_pc_im or runtime helper) has done so before we reach + * return from cpu_tb_exec. */ #define DISAS_EXIT DISAS_TARGET_9 diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_msr_i(DisasContext *s, uint32_t insn, gen_helper_msr_i_daifclear(cpu_env, t1); tcg_temp_free_i32(t1); /* For DAIFClear, exit the cpu loop to re-evaluate pending IRQs. */ - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; break; default: @@ -XXX,XX +XXX,XX @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread, if ((tb_cflags(s->base.tb) & CF_USE_ICOUNT) && (ri->type & ARM_CP_IO)) { /* I/O operations must end the TB here (whether read or write) */ - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; } if (!isread && !(ri->type & ARM_CP_SUPPRESS_TB_END)) { /* @@ -XXX,XX +XXX,XX @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread, * but allow this to be suppressed by the register definition * (usually only necessary to work around guest bugs). */ - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; } } @@ -XXX,XX +XXX,XX @@ static void aarch64_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) gen_goto_tb(dc, 1, dc->base.pc_next); break; default: - case DISAS_UPDATE: + case DISAS_UPDATE_EXIT: gen_a64_set_pc_im(dc->base.pc_next); /* fall through */ case DISAS_EXIT: diff --git a/target/arm/translate-vfp.inc.c b/target/arm/translate-vfp.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-vfp.inc.c +++ b/target/arm/translate-vfp.inc.c @@ -XXX,XX +XXX,XX @@ static bool full_vfp_access_check(DisasContext *s, bool ignore_vfp_enabled) * this to be the last insn in the TB). */ if (tb_cflags(s->base.tb) & CF_USE_ICOUNT) { - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; gen_io_start(); } gen_helper_v7m_preserve_fp_state(cpu_env); @@ -XXX,XX +XXX,XX @@ static bool trans_VLLDM_VLSTM(DisasContext *s, arg_VLLDM_VLSTM *a) tcg_temp_free_i32(fptr); /* End the TB, because we have updated FP control bits */ - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; return true; } 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 gen_msr_banked(DisasContext *s, int r, int sysm, int rn) tcg_temp_free_i32(tcg_tgtmode); tcg_temp_free_i32(tcg_regno); tcg_temp_free_i32(tcg_reg); - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; } static void gen_mrs_banked(DisasContext *s, int r, int sysm, int rn) @@ -XXX,XX +XXX,XX @@ static void gen_mrs_banked(DisasContext *s, int r, int sysm, int rn) tcg_temp_free_i32(tcg_tgtmode); tcg_temp_free_i32(tcg_regno); store_reg(s, rn, tcg_reg); - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; } /* Store value to PC as for an exception return (ie don't @@ -XXX,XX +XXX,XX @@ static void gen_srs(DisasContext *s, tcg_temp_free_i32(tmp); } tcg_temp_free_i32(addr); - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; } /* Generate a label used for skipping this instruction */ @@ -XXX,XX +XXX,XX @@ static bool trans_SETEND(DisasContext *s, arg_SETEND *a) } if (a->E != (s->be_data == MO_BE)) { gen_helper_setend(cpu_env); - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; } return true; } @@ -XXX,XX +XXX,XX @@ static void arm_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) break; case DISAS_NEXT: case DISAS_TOO_MANY: - case DISAS_UPDATE: + case DISAS_UPDATE_EXIT: gen_set_pc_im(dc, dc->base.pc_next); /* fall through */ default: @@ -XXX,XX +XXX,XX @@ static void arm_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) case DISAS_JUMP: gen_goto_ptr(); break; - case DISAS_UPDATE: + case DISAS_UPDATE_EXIT: gen_set_pc_im(dc, dc->base.pc_next); /* fall through */ default: -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Add an option that writes back the PC, like DISAS_UPDATE_EXIT, but does not exit back to the main loop. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-7-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.h | 2 ++ target/arm/translate-a64.c | 3 +++ target/arm/translate.c | 4 ++++ 3 files changed, 9 insertions(+) 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 void disas_set_insn_syndrome(DisasContext *s, uint32_t syn) * return from cpu_tb_exec. */ #define DISAS_EXIT DISAS_TARGET_9 +/* CPU state was modified dynamically; no need to exit, but do not chain. */ +#define DISAS_UPDATE_NOCHAIN DISAS_TARGET_10 #ifdef TARGET_AARCH64 void a64_translate_init(void); 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_tb_stop(DisasContextBase *dcbase, CPUState *cpu) case DISAS_EXIT: tcg_gen_exit_tb(NULL, 0); break; + case DISAS_UPDATE_NOCHAIN: + gen_a64_set_pc_im(dc->base.pc_next); + /* fall through */ case DISAS_JUMP: tcg_gen_lookup_and_goto_ptr(); break; 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_tb_stop(DisasContextBase *dcbase, CPUState *cpu) case DISAS_NEXT: case DISAS_TOO_MANY: case DISAS_UPDATE_EXIT: + case DISAS_UPDATE_NOCHAIN: gen_set_pc_im(dc, dc->base.pc_next); /* fall through */ default: @@ -XXX,XX +XXX,XX @@ static void arm_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) case DISAS_TOO_MANY: gen_goto_tb(dc, 1, dc->base.pc_next); break; + case DISAS_UPDATE_NOCHAIN: + gen_set_pc_im(dc, dc->base.pc_next); + /* fall through */ case DISAS_JUMP: gen_goto_ptr(); break; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> This is TFSRE0_EL1, TFSR_EL1, TFSR_EL2, TFSR_EL3, RGSR_EL1, GCR_EL1, GMID_EL1, and PSTATE.TCO. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-8-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 4 ++ target/arm/internals.h | 9 ++++ target/arm/helper.c | 94 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-a64.c | 21 +++++++++ 4 files changed, 128 insertions(+) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ typedef struct CPUARMState { uint64_t pmccfiltr_el0; /* Performance Monitor Filter Register */ uint64_t vpidr_el2; /* Virtualization Processor ID Register */ uint64_t vmpidr_el2; /* Virtualization Multiprocessor ID Register */ + uint64_t tfsr_el[4]; /* tfsre0_el1 is index 0. */ + uint64_t gcr_el1; + uint64_t rgsr_el1; } cp15; struct { @@ -XXX,XX +XXX,XX @@ void pmu_init(ARMCPU *cpu); #define PSTATE_SS (1U << 21) #define PSTATE_PAN (1U << 22) #define PSTATE_UAO (1U << 23) +#define PSTATE_TCO (1U << 25) #define PSTATE_V (1U << 28) #define PSTATE_C (1U << 29) #define PSTATE_Z (1U << 30) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ static inline uint32_t aarch64_pstate_valid_mask(const ARMISARegisters *id) if (isar_feature_aa64_uao(id)) { valid |= PSTATE_UAO; } + if (isar_feature_aa64_mte(id)) { + valid |= PSTATE_TCO; + } return valid; } @@ -XXX,XX +XXX,XX @@ void arm_log_exception(int idx); #endif /* !CONFIG_USER_ONLY */ +/* + * The log2 of the words in the tag block, for GMID_EL1.BS. + * The is the maximum, 256 bytes, which manipulates 64-bits of tags. + */ +#define GMID_EL1_BS 6 + #endif diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void define_arm_vh_e2h_redirects_aliases(ARMCPU *cpu) { K(3, 0, 1, 2, 0), K(3, 4, 1, 2, 0), K(3, 5, 1, 2, 0), "ZCR_EL1", "ZCR_EL2", "ZCR_EL12", isar_feature_aa64_sve }, + { K(3, 0, 5, 6, 0), K(3, 4, 5, 6, 0), K(3, 5, 5, 6, 0), + "TFSR_EL1", "TFSR_EL2", "TFSR_EL12", isar_feature_aa64_mte }, + /* TODO: ARMv8.2-SPE -- PMSCR_EL2 */ /* TODO: ARMv8.4-Trace -- TRFCR_EL2 */ }; @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo dcpodp_reg[] = { }; #endif /*CONFIG_USER_ONLY*/ +static CPAccessResult access_aa64_tid5(CPUARMState *env, const ARMCPRegInfo *ri, + bool isread) +{ + if ((arm_current_el(env) < 2) && (arm_hcr_el2_eff(env) & HCR_TID5)) { + return CP_ACCESS_TRAP_EL2; + } + + return CP_ACCESS_OK; +} + +static CPAccessResult access_mte(CPUARMState *env, const ARMCPRegInfo *ri, + bool isread) +{ + int el = arm_current_el(env); + + if (el < 2 && + arm_feature(env, ARM_FEATURE_EL2) && + !(arm_hcr_el2_eff(env) & HCR_ATA)) { + return CP_ACCESS_TRAP_EL2; + } + if (el < 3 && + arm_feature(env, ARM_FEATURE_EL3) && + !(env->cp15.scr_el3 & SCR_ATA)) { + return CP_ACCESS_TRAP_EL3; + } + return CP_ACCESS_OK; +} + +static uint64_t tco_read(CPUARMState *env, const ARMCPRegInfo *ri) +{ + return env->pstate & PSTATE_TCO; +} + +static void tco_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t val) +{ + env->pstate = (env->pstate & ~PSTATE_TCO) | (val & PSTATE_TCO); +} + +static const ARMCPRegInfo mte_reginfo[] = { + { .name = "TFSRE0_EL1", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 0, .crn = 5, .crm = 6, .opc2 = 1, + .access = PL1_RW, .accessfn = access_mte, + .fieldoffset = offsetof(CPUARMState, cp15.tfsr_el[0]) }, + { .name = "TFSR_EL1", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 0, .crn = 5, .crm = 6, .opc2 = 0, + .access = PL1_RW, .accessfn = access_mte, + .fieldoffset = offsetof(CPUARMState, cp15.tfsr_el[1]) }, + { .name = "TFSR_EL2", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 4, .crn = 5, .crm = 6, .opc2 = 0, + .access = PL2_RW, .accessfn = access_mte, + .fieldoffset = offsetof(CPUARMState, cp15.tfsr_el[2]) }, + { .name = "TFSR_EL3", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 6, .crn = 5, .crm = 6, .opc2 = 0, + .access = PL3_RW, + .fieldoffset = offsetof(CPUARMState, cp15.tfsr_el[3]) }, + { .name = "RGSR_EL1", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 0, .crn = 1, .crm = 0, .opc2 = 5, + .access = PL1_RW, .accessfn = access_mte, + .fieldoffset = offsetof(CPUARMState, cp15.rgsr_el1) }, + { .name = "GCR_EL1", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 0, .crn = 1, .crm = 0, .opc2 = 6, + .access = PL1_RW, .accessfn = access_mte, + .fieldoffset = offsetof(CPUARMState, cp15.gcr_el1) }, + { .name = "GMID_EL1", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 1, .crn = 0, .crm = 0, .opc2 = 4, + .access = PL1_R, .accessfn = access_aa64_tid5, + .type = ARM_CP_CONST, .resetvalue = GMID_EL1_BS }, + { .name = "TCO", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 3, .crn = 4, .crm = 2, .opc2 = 7, + .type = ARM_CP_NO_RAW, + .access = PL0_RW, .readfn = tco_read, .writefn = tco_write }, + REGINFO_SENTINEL +}; + +static const ARMCPRegInfo mte_tco_ro_reginfo[] = { + { .name = "TCO", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 3, .crn = 4, .crm = 2, .opc2 = 7, + .type = ARM_CP_CONST, .access = PL0_RW, }, + REGINFO_SENTINEL +}; #endif static CPAccessResult access_predinv(CPUARMState *env, const ARMCPRegInfo *ri, @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) } } #endif /*CONFIG_USER_ONLY*/ + + /* + * If full MTE is enabled, add all of the system registers. + * If only "instructions available at EL0" are enabled, + * then define only a RAZ/WI version of PSTATE.TCO. + */ + if (cpu_isar_feature(aa64_mte, cpu)) { + define_arm_cp_regs(cpu, mte_reginfo); + } else if (cpu_isar_feature(aa64_mte_insn_reg, cpu)) { + define_arm_cp_regs(cpu, mte_tco_ro_reginfo); + } #endif if (cpu_isar_feature(any_predinv, cpu)) { diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_msr_i(DisasContext *s, uint32_t insn, s->base.is_jmp = DISAS_UPDATE_EXIT; break; + case 0x1c: /* TCO */ + if (dc_isar_feature(aa64_mte, s)) { + /* Full MTE is enabled -- set the TCO bit as directed. */ + if (crm & 1) { + set_pstate_bits(PSTATE_TCO); + } else { + clear_pstate_bits(PSTATE_TCO); + } + t1 = tcg_const_i32(s->current_el); + gen_helper_rebuild_hflags_a64(cpu_env, t1); + tcg_temp_free_i32(t1); + /* Many factors, including TCO, go into MTE_ACTIVE. */ + s->base.is_jmp = DISAS_UPDATE_NOCHAIN; + } else if (dc_isar_feature(aa64_mte_insn_reg, s)) { + /* Only "instructions accessible at EL0" -- PSTATE.TCO is WI. */ + s->base.is_jmp = DISAS_NEXT; + } else { + goto do_unallocated; + } + break; + default: do_unallocated: unallocated_encoding(s); -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Cache the composite ATA setting. Cache when MTE is fully enabled, i.e. access to tags are enabled and tag checks affect the PE. Do this for both the normal context and the UNPRIV context. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-9-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 12 ++++++++---- target/arm/internals.h | 18 +++++++++++++++++ target/arm/translate.h | 5 +++++ target/arm/helper.c | 40 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-a64.c | 4 ++++ 5 files changed, 75 insertions(+), 4 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 @@ typedef ARMCPU ArchCPU; * | | | TBFLAG_A32 | | * | | +-----+----------+ TBFLAG_AM32 | * | TBFLAG_ANY | |TBFLAG_M32| | - * | | +-+----------+--------------| - * | | | TBFLAG_A64 | - * +--------------+---------+---------------------------+ - * 31 20 15 0 + * | +-----------+----------+--------------| + * | | TBFLAG_A64 | + * +--------------+-------------------------------------+ + * 31 20 0 * * Unless otherwise noted, these bits are cached in env->hflags. */ @@ -XXX,XX +XXX,XX @@ FIELD(TBFLAG_A64, BT, 9, 1) FIELD(TBFLAG_A64, BTYPE, 10, 2) /* Not cached. */ FIELD(TBFLAG_A64, TBID, 12, 2) FIELD(TBFLAG_A64, UNPRIV, 14, 1) +FIELD(TBFLAG_A64, ATA, 15, 1) +FIELD(TBFLAG_A64, TCMA, 16, 2) +FIELD(TBFLAG_A64, MTE_ACTIVE, 18, 1) +FIELD(TBFLAG_A64, MTE0_ACTIVE, 19, 1) /** * cpu_mmu_index: diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ static inline int exception_target_el(CPUARMState *env) return target_el; } +/* Determine if allocation tags are available. */ +static inline bool allocation_tag_access_enabled(CPUARMState *env, int el, + uint64_t sctlr) +{ + if (el < 3 + && arm_feature(env, ARM_FEATURE_EL3) + && !(env->cp15.scr_el3 & SCR_ATA)) { + return false; + } + if (el < 2 + && arm_feature(env, ARM_FEATURE_EL2) + && !(arm_hcr_el2_eff(env) & HCR_ATA)) { + return false; + } + sctlr &= (el == 0 ? SCTLR_ATA0 : SCTLR_ATA); + return sctlr != 0; +} + #ifndef CONFIG_USER_ONLY /* Security attributes for an address, as returned by v8m_security_lookup. */ 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 { ARMMMUIdx mmu_idx; /* MMU index to use for normal loads/stores */ uint8_t tbii; /* TBI1|TBI0 for insns */ uint8_t tbid; /* TBI1|TBI0 for data */ + uint8_t tcma; /* TCMA1|TCMA0 for MTE */ bool ns; /* Use non-secure CPREG bank on access */ int fp_excp_el; /* FP exception EL or 0 if enabled */ int sve_excp_el; /* SVE exception EL or 0 if enabled */ @@ -XXX,XX +XXX,XX @@ typedef struct DisasContext { bool unpriv; /* True if v8.3-PAuth is active. */ bool pauth_active; + /* True if v8.5-MTE access to tags is enabled. */ + bool ata; + /* True if v8.5-MTE tag checks affect the PE; index with is_unpriv. */ + bool mte_active[2]; /* True with v8.5-BTI and SCTLR_ELx.BT* set. */ bool bt; /* True if any CP15 access is trapped by HSTR_EL2 */ diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static int aa64_va_parameter_tbid(uint64_t tcr, ARMMMUIdx mmu_idx) } } +static int aa64_va_parameter_tcma(uint64_t tcr, ARMMMUIdx mmu_idx) +{ + if (regime_has_2_ranges(mmu_idx)) { + return extract64(tcr, 57, 2); + } else { + /* Replicate the single TCMA bit so we always have 2 bits. */ + return extract32(tcr, 30, 1) * 3; + } +} + ARMVAParameters aa64_va_parameters(CPUARMState *env, uint64_t va, ARMMMUIdx mmu_idx, bool data) { @@ -XXX,XX +XXX,XX @@ static uint32_t rebuild_hflags_a64(CPUARMState *env, int el, int fp_el, } } + if (cpu_isar_feature(aa64_mte, env_archcpu(env))) { + /* + * Set MTE_ACTIVE if any access may be Checked, and leave clear + * if all accesses must be Unchecked: + * 1) If no TBI, then there are no tags in the address to check, + * 2) If Tag Check Override, then all accesses are Unchecked, + * 3) If Tag Check Fail == 0, then Checked access have no effect, + * 4) If no Allocation Tag Access, then all accesses are Unchecked. + */ + if (allocation_tag_access_enabled(env, el, sctlr)) { + flags = FIELD_DP32(flags, TBFLAG_A64, ATA, 1); + if (tbid + && !(env->pstate & PSTATE_TCO) + && (sctlr & (el == 0 ? SCTLR_TCF0 : SCTLR_TCF))) { + flags = FIELD_DP32(flags, TBFLAG_A64, MTE_ACTIVE, 1); + } + } + /* And again for unprivileged accesses, if required. */ + if (FIELD_EX32(flags, TBFLAG_A64, UNPRIV) + && tbid + && !(env->pstate & PSTATE_TCO) + && (sctlr & SCTLR_TCF0) + && allocation_tag_access_enabled(env, 0, sctlr)) { + flags = FIELD_DP32(flags, TBFLAG_A64, MTE0_ACTIVE, 1); + } + /* Cache TCMA as well as TBI. */ + flags = FIELD_DP32(flags, TBFLAG_A64, TCMA, + aa64_va_parameter_tcma(tcr, mmu_idx)); + } + return rebuild_hflags_common(env, fp_el, mmu_idx, flags); } 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_init_disas_context(DisasContextBase *dcbase, dc->mmu_idx = core_to_aa64_mmu_idx(core_mmu_idx); dc->tbii = FIELD_EX32(tb_flags, TBFLAG_A64, TBII); dc->tbid = FIELD_EX32(tb_flags, TBFLAG_A64, TBID); + dc->tcma = FIELD_EX32(tb_flags, TBFLAG_A64, TCMA); dc->current_el = arm_mmu_idx_to_el(dc->mmu_idx); #if !defined(CONFIG_USER_ONLY) dc->user = (dc->current_el == 0); @@ -XXX,XX +XXX,XX @@ static void aarch64_tr_init_disas_context(DisasContextBase *dcbase, dc->bt = FIELD_EX32(tb_flags, TBFLAG_A64, BT); dc->btype = FIELD_EX32(tb_flags, TBFLAG_A64, BTYPE); dc->unpriv = FIELD_EX32(tb_flags, TBFLAG_A64, UNPRIV); + dc->ata = FIELD_EX32(tb_flags, TBFLAG_A64, ATA); + dc->mte_active[0] = FIELD_EX32(tb_flags, TBFLAG_A64, MTE_ACTIVE); + dc->mte_active[1] = FIELD_EX32(tb_flags, TBFLAG_A64, MTE0_ACTIVE); dc->vec_len = 0; dc->vec_stride = 0; dc->cp_regs = arm_cpu->cp_regs; -- 2.20.1 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: 20200626033144.790098-10-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 2 ++ target/arm/internals.h | 5 +++ target/arm/mte_helper.c | 72 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-a64.c | 18 ++++++++++ target/arm/Makefile.objs | 1 + 5 files changed, 98 insertions(+) create mode 100644 target/arm/mte_helper.c diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(autda, TCG_CALL_NO_WG, i64, env, i64, i64) DEF_HELPER_FLAGS_3(autdb, TCG_CALL_NO_WG, i64, env, i64, i64) DEF_HELPER_FLAGS_2(xpaci, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_2(xpacd, TCG_CALL_NO_RWG_SE, i64, env, i64) + +DEF_HELPER_FLAGS_3(irg, TCG_CALL_NO_RWG, i64, env, i64, i64) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ void arm_log_exception(int idx); */ #define GMID_EL1_BS 6 +static inline uint64_t address_with_allocation_tag(uint64_t ptr, int rtag) +{ + return deposit64(ptr, 56, 4, rtag); +} + #endif diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ +/* + * ARM v8.5-MemTag Operations + * + * Copyright (c) 2020 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 "internals.h" +#include "exec/exec-all.h" +#include "exec/cpu_ldst.h" +#include "exec/helper-proto.h" + + +static int choose_nonexcluded_tag(int tag, int offset, uint16_t exclude) +{ + if (exclude == 0xffff) { + return 0; + } + if (offset == 0) { + while (exclude & (1 << tag)) { + tag = (tag + 1) & 15; + } + } else { + do { + do { + tag = (tag + 1) & 15; + } while (exclude & (1 << tag)); + } while (--offset > 0); + } + return tag; +} + +uint64_t HELPER(irg)(CPUARMState *env, uint64_t rn, uint64_t rm) +{ + int rtag; + + /* + * Our IMPDEF choice for GCR_EL1.RRND==1 is to behave as if + * GCR_EL1.RRND==0, always producing deterministic results. + */ + uint16_t exclude = extract32(rm | env->cp15.gcr_el1, 0, 16); + int start = extract32(env->cp15.rgsr_el1, 0, 4); + int seed = extract32(env->cp15.rgsr_el1, 8, 16); + int offset, i; + + /* RandomTag */ + for (i = offset = 0; i < 4; ++i) { + /* NextRandomTagBit */ + int top = (extract32(seed, 5, 1) ^ extract32(seed, 3, 1) ^ + extract32(seed, 2, 1) ^ extract32(seed, 0, 1)); + seed = (top << 15) | (seed >> 1); + offset |= top << i; + } + rtag = choose_nonexcluded_tag(start, offset, exclude); + env->cp15.rgsr_el1 = rtag | (seed << 8); + + return address_with_allocation_tag(rn, rtag); +} 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 TCGv_i64 clean_data_tbi(DisasContext *s, TCGv_i64 addr) return clean; } +/* Insert a zero tag into src, with the result at dst. */ +static void gen_address_with_allocation_tag0(TCGv_i64 dst, TCGv_i64 src) +{ + tcg_gen_andi_i64(dst, src, ~MAKE_64BIT_MASK(56, 4)); +} + typedef struct DisasCompare64 { TCGCond cond; TCGv_i64 value; @@ -XXX,XX +XXX,XX @@ static void disas_data_proc_2src(DisasContext *s, uint32_t insn) case 3: /* SDIV */ handle_div(s, true, sf, rm, rn, rd); break; + case 4: /* IRG */ + if (sf == 0 || !dc_isar_feature(aa64_mte_insn_reg, s)) { + goto do_unallocated; + } + if (s->ata) { + gen_helper_irg(cpu_reg_sp(s, rd), cpu_env, + cpu_reg_sp(s, rn), cpu_reg(s, rm)); + } else { + gen_address_with_allocation_tag0(cpu_reg_sp(s, rd), + cpu_reg_sp(s, rn)); + } + break; case 8: /* LSLV */ handle_shift_reg(s, A64_SHIFT_TYPE_LSL, sf, rm, rn, rd); break; diff --git a/target/arm/Makefile.objs b/target/arm/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/target/arm/Makefile.objs +++ b/target/arm/Makefile.objs @@ -XXX,XX +XXX,XX @@ obj-$(CONFIG_SOFTMMU) += psci.o obj-$(TARGET_AARCH64) += translate-a64.o helper-a64.o obj-$(TARGET_AARCH64) += translate-sve.o sve_helper.o obj-$(TARGET_AARCH64) += pauth_helper.o +obj-$(TARGET_AARCH64) += mte_helper.o -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> The current Arm ARM has adjusted the official decode of "Add/subtract (immediate)" so that the shift field is only bit 22, and bit 23 is part of the op1 field of the parent category "Data processing - immediate". Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-11-richard.henderson@linaro.org Suggested-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 23 ++++++++--------------- 1 file changed, 8 insertions(+), 15 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_pc_rel_adr(DisasContext *s, uint32_t insn) /* * Add/subtract (immediate) * - * 31 30 29 28 24 23 22 21 10 9 5 4 0 - * +--+--+--+-----------+-----+-------------+-----+-----+ - * |sf|op| S| 1 0 0 0 1 |shift| imm12 | Rn | Rd | - * +--+--+--+-----------+-----+-------------+-----+-----+ + * 31 30 29 28 23 22 21 10 9 5 4 0 + * +--+--+--+-------------+--+-------------+-----+-----+ + * |sf|op| S| 1 0 0 0 1 0 |sh| imm12 | Rn | Rd | + * +--+--+--+-------------+--+-------------+-----+-----+ * * sf: 0 -> 32bit, 1 -> 64bit * op: 0 -> add , 1 -> sub * S: 1 -> set flags - * shift: 00 -> LSL imm by 0, 01 -> LSL imm by 12 + * sh: 1 -> LSL imm by 12 */ static void disas_add_sub_imm(DisasContext *s, uint32_t insn) { int rd = extract32(insn, 0, 5); int rn = extract32(insn, 5, 5); uint64_t imm = extract32(insn, 10, 12); - int shift = extract32(insn, 22, 2); + bool shift = extract32(insn, 22, 1); bool setflags = extract32(insn, 29, 1); bool sub_op = extract32(insn, 30, 1); bool is_64bit = extract32(insn, 31, 1); @@ -XXX,XX +XXX,XX @@ static void disas_add_sub_imm(DisasContext *s, uint32_t insn) TCGv_i64 tcg_rd = setflags ? cpu_reg(s, rd) : cpu_reg_sp(s, rd); TCGv_i64 tcg_result; - switch (shift) { - case 0x0: - break; - case 0x1: + if (shift) { imm <<= 12; - break; - default: - unallocated_encoding(s); - return; } tcg_result = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ static void disas_data_proc_imm(DisasContext *s, uint32_t insn) case 0x20: case 0x21: /* PC-rel. addressing */ disas_pc_rel_adr(s, insn); break; - case 0x22: case 0x23: /* Add/subtract (immediate) */ + case 0x22: /* Add/subtract (immediate) */ disas_add_sub_imm(s, insn); break; case 0x24: /* Logical (immediate) */ -- 2.20.1 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: 20200626033144.790098-12-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 1 + target/arm/internals.h | 9 +++++++ target/arm/mte_helper.c | 10 ++++++++ target/arm/translate-a64.c | 51 ++++++++++++++++++++++++++++++++++++++ 4 files changed, 71 insertions(+) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_2(xpaci, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_2(xpacd, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_3(irg, TCG_CALL_NO_RWG, i64, env, i64, i64) +DEF_HELPER_FLAGS_4(addsubg, TCG_CALL_NO_RWG_SE, i64, env, i64, s32, i32) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ void arm_log_exception(int idx); */ #define GMID_EL1_BS 6 +/* We associate one allocation tag per 16 bytes, the minimum. */ +#define LOG2_TAG_GRANULE 4 +#define TAG_GRANULE (1 << LOG2_TAG_GRANULE) + +static inline int allocation_tag_from_addr(uint64_t ptr) +{ + return extract64(ptr, 56, 4); +} + static inline uint64_t address_with_allocation_tag(uint64_t ptr, int rtag) { return deposit64(ptr, 56, 4, rtag); diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(irg)(CPUARMState *env, uint64_t rn, uint64_t rm) return address_with_allocation_tag(rn, rtag); } + +uint64_t HELPER(addsubg)(CPUARMState *env, uint64_t ptr, + int32_t offset, uint32_t tag_offset) +{ + int start_tag = allocation_tag_from_addr(ptr); + uint16_t exclude = extract32(env->cp15.gcr_el1, 0, 16); + int rtag = choose_nonexcluded_tag(start_tag, tag_offset, exclude); + + return address_with_allocation_tag(ptr + offset, rtag); +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_add_sub_imm(DisasContext *s, uint32_t insn) tcg_temp_free_i64(tcg_result); } +/* + * Add/subtract (immediate, with tags) + * + * 31 30 29 28 23 22 21 16 14 10 9 5 4 0 + * +--+--+--+-------------+--+---------+--+-------+-----+-----+ + * |sf|op| S| 1 0 0 0 1 1 |o2| uimm6 |o3| uimm4 | Rn | Rd | + * +--+--+--+-------------+--+---------+--+-------+-----+-----+ + * + * op: 0 -> add, 1 -> sub + */ +static void disas_add_sub_imm_with_tags(DisasContext *s, uint32_t insn) +{ + int rd = extract32(insn, 0, 5); + int rn = extract32(insn, 5, 5); + int uimm4 = extract32(insn, 10, 4); + int uimm6 = extract32(insn, 16, 6); + bool sub_op = extract32(insn, 30, 1); + TCGv_i64 tcg_rn, tcg_rd; + int imm; + + /* Test all of sf=1, S=0, o2=0, o3=0. */ + if ((insn & 0xa040c000u) != 0x80000000u || + !dc_isar_feature(aa64_mte_insn_reg, s)) { + unallocated_encoding(s); + return; + } + + imm = uimm6 << LOG2_TAG_GRANULE; + if (sub_op) { + imm = -imm; + } + + tcg_rn = cpu_reg_sp(s, rn); + tcg_rd = cpu_reg_sp(s, rd); + + if (s->ata) { + TCGv_i32 offset = tcg_const_i32(imm); + TCGv_i32 tag_offset = tcg_const_i32(uimm4); + + gen_helper_addsubg(tcg_rd, cpu_env, tcg_rn, offset, tag_offset); + tcg_temp_free_i32(tag_offset); + tcg_temp_free_i32(offset); + } else { + tcg_gen_addi_i64(tcg_rd, tcg_rn, imm); + gen_address_with_allocation_tag0(tcg_rd, tcg_rd); + } +} + /* The input should be a value in the bottom e bits (with higher * bits zero); returns that value replicated into every element * of size e in a 64 bit integer. @@ -XXX,XX +XXX,XX @@ static void disas_data_proc_imm(DisasContext *s, uint32_t insn) case 0x22: /* Add/subtract (immediate) */ disas_add_sub_imm(s, insn); break; + case 0x23: /* Add/subtract (immediate, with tags) */ + disas_add_sub_imm_with_tags(s, insn); + break; case 0x24: /* Logical (immediate) */ disas_logic_imm(s, insn); break; -- 2.20.1 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: 20200626033144.790098-13-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_data_proc_2src(DisasContext *s, uint32_t insn) cpu_reg_sp(s, rn)); } break; + case 5: /* GMI */ + if (sf == 0 || !dc_isar_feature(aa64_mte_insn_reg, s)) { + goto do_unallocated; + } else { + TCGv_i64 t1 = tcg_const_i64(1); + TCGv_i64 t2 = tcg_temp_new_i64(); + + tcg_gen_extract_i64(t2, cpu_reg_sp(s, rn), 56, 4); + tcg_gen_shl_i64(t1, t1, t2); + tcg_gen_or_i64(cpu_reg(s, rd), cpu_reg(s, rm), t1); + + tcg_temp_free_i64(t1); + tcg_temp_free_i64(t2); + } + break; case 8: /* LSLV */ handle_shift_reg(s, A64_SHIFT_TYPE_LSL, sf, rm, rn, rd); break; -- 2.20.1 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: 20200626033144.790098-14-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 24 ++++++++++++++++++++++-- 1 file changed, 22 insertions(+), 2 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_crc32(DisasContext *s, */ static void disas_data_proc_2src(DisasContext *s, uint32_t insn) { - unsigned int sf, rm, opcode, rn, rd; + unsigned int sf, rm, opcode, rn, rd, setflag; sf = extract32(insn, 31, 1); + setflag = extract32(insn, 29, 1); rm = extract32(insn, 16, 5); opcode = extract32(insn, 10, 6); rn = extract32(insn, 5, 5); rd = extract32(insn, 0, 5); - if (extract32(insn, 29, 1)) { + if (setflag && opcode != 0) { unallocated_encoding(s); return; } switch (opcode) { + case 0: /* SUBP(S) */ + if (sf == 0 || !dc_isar_feature(aa64_mte_insn_reg, s)) { + goto do_unallocated; + } else { + TCGv_i64 tcg_n, tcg_m, tcg_d; + + tcg_n = read_cpu_reg_sp(s, rn, true); + tcg_m = read_cpu_reg_sp(s, rm, true); + tcg_gen_sextract_i64(tcg_n, tcg_n, 0, 56); + tcg_gen_sextract_i64(tcg_m, tcg_m, 0, 56); + tcg_d = cpu_reg(s, rd); + + if (setflag) { + gen_sub_CC(true, tcg_d, tcg_n, tcg_m); + } else { + tcg_gen_sub_i64(tcg_d, tcg_n, tcg_m); + } + } + break; case 2: /* UDIV */ handle_div(s, false, sf, rm, rn, rd); break; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Use the same code as system mode, so that we generate the same exception + syndrome for the unaligned access. For the moment, if MTE is enabled so that this path is reachable, this would generate a SIGSEGV in the user-only cpu_loop. Decoding the syndrome to produce the proper SIGBUS will be done later. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-15-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 2 +- target/arm/tlb_helper.c | 4 ++-- 2 files changed, 3 insertions(+), 3 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_class_init(ObjectClass *oc, void *data) cc->tlb_fill = arm_cpu_tlb_fill; cc->debug_excp_handler = arm_debug_excp_handler; cc->debug_check_watchpoint = arm_debug_check_watchpoint; -#if !defined(CONFIG_USER_ONLY) cc->do_unaligned_access = arm_cpu_do_unaligned_access; +#if !defined(CONFIG_USER_ONLY) cc->do_transaction_failed = arm_cpu_do_transaction_failed; cc->adjust_watchpoint_address = arm_adjust_watchpoint_address; #endif /* CONFIG_TCG && !CONFIG_USER_ONLY */ diff --git a/target/arm/tlb_helper.c b/target/arm/tlb_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tlb_helper.c +++ b/target/arm/tlb_helper.c @@ -XXX,XX +XXX,XX @@ #include "internals.h" #include "exec/exec-all.h" -#if !defined(CONFIG_USER_ONLY) - static inline uint32_t merge_syn_data_abort(uint32_t template_syn, unsigned int target_el, bool same_el, bool ea, @@ -XXX,XX +XXX,XX @@ void arm_cpu_do_unaligned_access(CPUState *cs, vaddr vaddr, arm_deliver_fault(cpu, vaddr, access_type, mmu_idx, &fi); } +#if !defined(CONFIG_USER_ONLY) + /* * arm_cpu_do_transaction_failed: handle a memory system error response * (eg "no device/memory present at address") by raising an external abort -- 2.20.1 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: 20200626033144.790098-16-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 7 ++ target/arm/helper.h | 2 + target/arm/mte_helper.c | 194 +++++++++++++++++++++++++++++++++++++ target/arm/op_helper.c | 16 +++ target/arm/translate-a64.c | 172 +++++++++++++++++++++++++++++++- 5 files changed, 386 insertions(+), 5 deletions(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_2(xpacd, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_3(irg, TCG_CALL_NO_RWG, i64, env, i64, i64) DEF_HELPER_FLAGS_4(addsubg, TCG_CALL_NO_RWG_SE, i64, env, i64, s32, i32) +DEF_HELPER_FLAGS_3(ldg, TCG_CALL_NO_WG, i64, env, i64, i64) +DEF_HELPER_FLAGS_3(stg, TCG_CALL_NO_WG, void, env, i64, i64) +DEF_HELPER_FLAGS_3(stg_parallel, TCG_CALL_NO_WG, void, env, i64, i64) +DEF_HELPER_FLAGS_2(stg_stub, TCG_CALL_NO_WG, void, env, i64) +DEF_HELPER_FLAGS_3(st2g, TCG_CALL_NO_WG, void, env, i64, i64) +DEF_HELPER_FLAGS_3(st2g_parallel, TCG_CALL_NO_WG, void, env, i64, i64) +DEF_HELPER_FLAGS_2(st2g_stub, TCG_CALL_NO_WG, void, env, i64) 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_1(rebuild_hflags_a32_newel, TCG_CALL_NO_RWG, void, env) DEF_HELPER_FLAGS_2(rebuild_hflags_a32, TCG_CALL_NO_RWG, void, env, int) DEF_HELPER_FLAGS_2(rebuild_hflags_a64, TCG_CALL_NO_RWG, void, env, int) +DEF_HELPER_FLAGS_5(probe_access, TCG_CALL_NO_WG, void, env, tl, i32, i32, i32) + DEF_HELPER_1(vfp_get_fpscr, i32, env) DEF_HELPER_2(vfp_set_fpscr, void, env, i32) diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ static int choose_nonexcluded_tag(int tag, int offset, uint16_t exclude) return tag; } +/** + * allocation_tag_mem: + * @env: the cpu environment + * @ptr_mmu_idx: the addressing regime to use for the virtual address + * @ptr: the virtual address for which to look up tag memory + * @ptr_access: the access to use for the virtual address + * @ptr_size: the number of bytes in the normal memory access + * @tag_access: the access to use for the tag memory + * @tag_size: the number of bytes in the tag memory access + * @ra: the return address for exception handling + * + * Our tag memory is formatted as a sequence of little-endian nibbles. + * That is, the byte at (addr >> (LOG2_TAG_GRANULE + 1)) contains two + * tags, with the tag at [3:0] for the lower addr and the tag at [7:4] + * for the higher addr. + * + * Here, resolve the physical address from the virtual address, and return + * a pointer to the corresponding tag byte. Exit with exception if the + * virtual address is not accessible for @ptr_access. + * + * The @ptr_size and @tag_size values may not have an obvious relation + * due to the alignment of @ptr, and the number of tag checks required. + * + * If there is no tag storage corresponding to @ptr, return NULL. + */ +static uint8_t *allocation_tag_mem(CPUARMState *env, int ptr_mmu_idx, + uint64_t ptr, MMUAccessType ptr_access, + int ptr_size, MMUAccessType tag_access, + int tag_size, uintptr_t ra) +{ + /* Tag storage not implemented. */ + return NULL; +} + uint64_t HELPER(irg)(CPUARMState *env, uint64_t rn, uint64_t rm) { int rtag; @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(addsubg)(CPUARMState *env, uint64_t ptr, return address_with_allocation_tag(ptr + offset, rtag); } + +static int load_tag1(uint64_t ptr, uint8_t *mem) +{ + int ofs = extract32(ptr, LOG2_TAG_GRANULE, 1) * 4; + return extract32(*mem, ofs, 4); +} + +uint64_t HELPER(ldg)(CPUARMState *env, uint64_t ptr, uint64_t xt) +{ + int mmu_idx = cpu_mmu_index(env, false); + uint8_t *mem; + int rtag = 0; + + /* Trap if accessing an invalid page. */ + mem = allocation_tag_mem(env, mmu_idx, ptr, MMU_DATA_LOAD, 1, + MMU_DATA_LOAD, 1, GETPC()); + + /* Load if page supports tags. */ + if (mem) { + rtag = load_tag1(ptr, mem); + } + + return address_with_allocation_tag(xt, rtag); +} + +static void check_tag_aligned(CPUARMState *env, uint64_t ptr, uintptr_t ra) +{ + if (unlikely(!QEMU_IS_ALIGNED(ptr, TAG_GRANULE))) { + arm_cpu_do_unaligned_access(env_cpu(env), ptr, MMU_DATA_STORE, + cpu_mmu_index(env, false), ra); + g_assert_not_reached(); + } +} + +/* For use in a non-parallel context, store to the given nibble. */ +static void store_tag1(uint64_t ptr, uint8_t *mem, int tag) +{ + int ofs = extract32(ptr, LOG2_TAG_GRANULE, 1) * 4; + *mem = deposit32(*mem, ofs, 4, tag); +} + +/* For use in a parallel context, atomically store to the given nibble. */ +static void store_tag1_parallel(uint64_t ptr, uint8_t *mem, int tag) +{ + int ofs = extract32(ptr, LOG2_TAG_GRANULE, 1) * 4; + uint8_t old = atomic_read(mem); + + while (1) { + uint8_t new = deposit32(old, ofs, 4, tag); + uint8_t cmp = atomic_cmpxchg(mem, old, new); + if (likely(cmp == old)) { + return; + } + old = cmp; + } +} + +typedef void stg_store1(uint64_t, uint8_t *, int); + +static inline void do_stg(CPUARMState *env, uint64_t ptr, uint64_t xt, + uintptr_t ra, stg_store1 store1) +{ + int mmu_idx = cpu_mmu_index(env, false); + uint8_t *mem; + + check_tag_aligned(env, ptr, ra); + + /* Trap if accessing an invalid page. */ + mem = allocation_tag_mem(env, mmu_idx, ptr, MMU_DATA_STORE, TAG_GRANULE, + MMU_DATA_STORE, 1, ra); + + /* Store if page supports tags. */ + if (mem) { + store1(ptr, mem, allocation_tag_from_addr(xt)); + } +} + +void HELPER(stg)(CPUARMState *env, uint64_t ptr, uint64_t xt) +{ + do_stg(env, ptr, xt, GETPC(), store_tag1); +} + +void HELPER(stg_parallel)(CPUARMState *env, uint64_t ptr, uint64_t xt) +{ + do_stg(env, ptr, xt, GETPC(), store_tag1_parallel); +} + +void HELPER(stg_stub)(CPUARMState *env, uint64_t ptr) +{ + int mmu_idx = cpu_mmu_index(env, false); + uintptr_t ra = GETPC(); + + check_tag_aligned(env, ptr, ra); + probe_write(env, ptr, TAG_GRANULE, mmu_idx, ra); +} + +static inline void do_st2g(CPUARMState *env, uint64_t ptr, uint64_t xt, + uintptr_t ra, stg_store1 store1) +{ + int mmu_idx = cpu_mmu_index(env, false); + int tag = allocation_tag_from_addr(xt); + uint8_t *mem1, *mem2; + + check_tag_aligned(env, ptr, ra); + + /* + * Trap if accessing an invalid page(s). + * This takes priority over !allocation_tag_access_enabled. + */ + if (ptr & TAG_GRANULE) { + /* Two stores unaligned mod TAG_GRANULE*2 -- modify two bytes. */ + mem1 = allocation_tag_mem(env, mmu_idx, ptr, MMU_DATA_STORE, + TAG_GRANULE, MMU_DATA_STORE, 1, ra); + mem2 = allocation_tag_mem(env, mmu_idx, ptr + TAG_GRANULE, + MMU_DATA_STORE, TAG_GRANULE, + MMU_DATA_STORE, 1, ra); + + /* Store if page(s) support tags. */ + if (mem1) { + store1(TAG_GRANULE, mem1, tag); + } + if (mem2) { + store1(0, mem2, tag); + } + } else { + /* Two stores aligned mod TAG_GRANULE*2 -- modify one byte. */ + mem1 = allocation_tag_mem(env, mmu_idx, ptr, MMU_DATA_STORE, + 2 * TAG_GRANULE, MMU_DATA_STORE, 1, ra); + if (mem1) { + tag |= tag << 4; + atomic_set(mem1, tag); + } + } +} + +void HELPER(st2g)(CPUARMState *env, uint64_t ptr, uint64_t xt) +{ + do_st2g(env, ptr, xt, GETPC(), store_tag1); +} + +void HELPER(st2g_parallel)(CPUARMState *env, uint64_t ptr, uint64_t xt) +{ + do_st2g(env, ptr, xt, GETPC(), store_tag1_parallel); +} + +void HELPER(st2g_stub)(CPUARMState *env, uint64_t ptr) +{ + int mmu_idx = cpu_mmu_index(env, false); + uintptr_t ra = GETPC(); + int in_page = -(ptr | TARGET_PAGE_MASK); + + check_tag_aligned(env, ptr, ra); + + if (likely(in_page >= 2 * TAG_GRANULE)) { + probe_write(env, ptr, 2 * TAG_GRANULE, mmu_idx, ra); + } else { + probe_write(env, ptr, TAG_GRANULE, mmu_idx, ra); + probe_write(env, ptr + TAG_GRANULE, TAG_GRANULE, mmu_idx, ra); + } +} diff --git a/target/arm/op_helper.c b/target/arm/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/op_helper.c +++ b/target/arm/op_helper.c @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(ror_cc)(CPUARMState *env, uint32_t x, uint32_t i) return ((uint32_t)x >> shift) | (x << (32 - shift)); } } + +void HELPER(probe_access)(CPUARMState *env, target_ulong ptr, + uint32_t access_type, uint32_t mmu_idx, + uint32_t size) +{ + uint32_t in_page = -((uint32_t)ptr | TARGET_PAGE_SIZE); + uintptr_t ra = GETPC(); + + if (likely(size <= in_page)) { + probe_access(env, ptr, size, access_type, mmu_idx, ra); + } else { + probe_access(env, ptr, in_page, access_type, mmu_idx, ra); + probe_access(env, ptr + in_page, size - in_page, + access_type, mmu_idx, ra); + } +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void gen_address_with_allocation_tag0(TCGv_i64 dst, TCGv_i64 src) tcg_gen_andi_i64(dst, src, ~MAKE_64BIT_MASK(56, 4)); } +static void gen_probe_access(DisasContext *s, TCGv_i64 ptr, + MMUAccessType acc, int log2_size) +{ + TCGv_i32 t_acc = tcg_const_i32(acc); + TCGv_i32 t_idx = tcg_const_i32(get_mem_index(s)); + TCGv_i32 t_size = tcg_const_i32(1 << log2_size); + + gen_helper_probe_access(cpu_env, ptr, t_acc, t_idx, t_size); + tcg_temp_free_i32(t_acc); + tcg_temp_free_i32(t_idx); + tcg_temp_free_i32(t_size); +} + typedef struct DisasCompare64 { TCGCond cond; TCGv_i64 value; @@ -XXX,XX +XXX,XX @@ static void disas_ldst_single_struct(DisasContext *s, uint32_t insn) } } +/* + * Load/Store memory tags + * + * 31 30 29 24 22 21 12 10 5 0 + * +-----+-------------+-----+---+------+-----+------+------+ + * | 1 1 | 0 1 1 0 0 1 | op1 | 1 | imm9 | op2 | Rn | Rt | + * +-----+-------------+-----+---+------+-----+------+------+ + */ +static void disas_ldst_tag(DisasContext *s, uint32_t insn) +{ + int rt = extract32(insn, 0, 5); + int rn = extract32(insn, 5, 5); + uint64_t offset = sextract64(insn, 12, 9) << LOG2_TAG_GRANULE; + int op2 = extract32(insn, 10, 2); + int op1 = extract32(insn, 22, 2); + bool is_load = false, is_pair = false, is_zero = false; + int index = 0; + TCGv_i64 addr, clean_addr, tcg_rt; + + /* We checked insn bits [29:24,21] in the caller. */ + if (extract32(insn, 30, 2) != 3) { + goto do_unallocated; + } + + /* + * @index is a tri-state variable which has 3 states: + * < 0 : post-index, writeback + * = 0 : signed offset + * > 0 : pre-index, writeback + */ + switch (op1) { + case 0: + if (op2 != 0) { + /* STG */ + index = op2 - 2; + break; + } + goto do_unallocated; + case 1: + if (op2 != 0) { + /* STZG */ + is_zero = true; + index = op2 - 2; + } else { + /* LDG */ + is_load = true; + } + break; + case 2: + if (op2 != 0) { + /* ST2G */ + is_pair = true; + index = op2 - 2; + break; + } + goto do_unallocated; + case 3: + if (op2 != 0) { + /* STZ2G */ + is_pair = is_zero = true; + index = op2 - 2; + break; + } + goto do_unallocated; + + default: + do_unallocated: + unallocated_encoding(s); + return; + } + + if (!dc_isar_feature(aa64_mte_insn_reg, s)) { + goto do_unallocated; + } + + if (rn == 31) { + gen_check_sp_alignment(s); + } + + addr = read_cpu_reg_sp(s, rn, true); + if (index >= 0) { + /* pre-index or signed offset */ + tcg_gen_addi_i64(addr, addr, offset); + } + + if (is_load) { + tcg_gen_andi_i64(addr, addr, -TAG_GRANULE); + tcg_rt = cpu_reg(s, rt); + if (s->ata) { + gen_helper_ldg(tcg_rt, cpu_env, addr, tcg_rt); + } else { + clean_addr = clean_data_tbi(s, addr); + gen_probe_access(s, clean_addr, MMU_DATA_LOAD, MO_8); + gen_address_with_allocation_tag0(tcg_rt, addr); + } + } else { + tcg_rt = cpu_reg_sp(s, rt); + if (!s->ata) { + /* + * For STG and ST2G, we need to check alignment and probe memory. + * TODO: For STZG and STZ2G, we could rely on the stores below, + * at least for system mode; user-only won't enforce alignment. + */ + if (is_pair) { + gen_helper_st2g_stub(cpu_env, addr); + } else { + gen_helper_stg_stub(cpu_env, addr); + } + } else if (tb_cflags(s->base.tb) & CF_PARALLEL) { + if (is_pair) { + gen_helper_st2g_parallel(cpu_env, addr, tcg_rt); + } else { + gen_helper_stg_parallel(cpu_env, addr, tcg_rt); + } + } else { + if (is_pair) { + gen_helper_st2g(cpu_env, addr, tcg_rt); + } else { + gen_helper_stg(cpu_env, addr, tcg_rt); + } + } + } + + if (is_zero) { + TCGv_i64 clean_addr = clean_data_tbi(s, addr); + TCGv_i64 tcg_zero = tcg_const_i64(0); + int mem_index = get_mem_index(s); + int i, n = (1 + is_pair) << LOG2_TAG_GRANULE; + + tcg_gen_qemu_st_i64(tcg_zero, clean_addr, mem_index, + MO_Q | MO_ALIGN_16); + for (i = 8; i < n; i += 8) { + tcg_gen_addi_i64(clean_addr, clean_addr, 8); + tcg_gen_qemu_st_i64(tcg_zero, clean_addr, mem_index, MO_Q); + } + tcg_temp_free_i64(tcg_zero); + } + + if (index != 0) { + /* pre-index or post-index */ + if (index < 0) { + /* post-index */ + tcg_gen_addi_i64(addr, addr, offset); + } + tcg_gen_mov_i64(cpu_reg_sp(s, rn), addr); + } +} + /* Loads and stores */ static void disas_ldst(DisasContext *s, uint32_t insn) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst(DisasContext *s, uint32_t insn) case 0x0d: /* AdvSIMD load/store single structure */ disas_ldst_single_struct(s, insn); break; - case 0x19: /* LDAPR/STLR (unscaled immediate) */ - if (extract32(insn, 10, 2) != 0 || - extract32(insn, 21, 1) != 0) { + case 0x19: + if (extract32(insn, 21, 1) != 0) { + disas_ldst_tag(s, insn); + } else if (extract32(insn, 10, 2) == 0) { + disas_ldst_ldapr_stlr(s, insn); + } else { unallocated_encoding(s); - break; } - disas_ldst_ldapr_stlr(s, insn); break; default: unallocated_encoding(s); -- 2.20.1 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: 20200626033144.790098-17-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 29 ++++++++++++++++++++++++++--- 1 file changed, 26 insertions(+), 3 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_ld_lit(DisasContext *s, uint32_t insn) * +-----+-------+---+---+-------+---+-------+-------+------+------+ * * opc: LDP/STP/LDNP/STNP 00 -> 32 bit, 10 -> 64 bit - * LDPSW 01 + * LDPSW/STGP 01 * LDP/STP/LDNP/STNP (SIMD) 00 -> 32 bit, 01 -> 64 bit, 10 -> 128 bit * V: 0 -> GPR, 1 -> Vector * idx: 00 -> signed offset with non-temporal hint, 01 -> post-index, @@ -XXX,XX +XXX,XX @@ static void disas_ldst_pair(DisasContext *s, uint32_t insn) bool is_signed = false; bool postindex = false; bool wback = false; + bool set_tag = false; TCGv_i64 clean_addr, dirty_addr; @@ -XXX,XX +XXX,XX @@ static void disas_ldst_pair(DisasContext *s, uint32_t insn) if (is_vector) { size = 2 + opc; + } else if (opc == 1 && !is_load) { + /* STGP */ + if (!dc_isar_feature(aa64_mte_insn_reg, s) || index == 0) { + unallocated_encoding(s); + return; + } + size = 3; + set_tag = true; } else { size = 2 + extract32(opc, 1, 1); is_signed = extract32(opc, 0, 1); @@ -XXX,XX +XXX,XX @@ static void disas_ldst_pair(DisasContext *s, uint32_t insn) return; } - offset <<= size; + offset <<= (set_tag ? LOG2_TAG_GRANULE : size); if (rn == 31) { gen_check_sp_alignment(s); @@ -XXX,XX +XXX,XX @@ static void disas_ldst_pair(DisasContext *s, uint32_t insn) if (!postindex) { tcg_gen_addi_i64(dirty_addr, dirty_addr, offset); } - clean_addr = clean_data_tbi(s, dirty_addr); + if (set_tag) { + if (!s->ata) { + /* + * TODO: We could rely on the stores below, at least for + * system mode, if we arrange to add MO_ALIGN_16. + */ + gen_helper_stg_stub(cpu_env, dirty_addr); + } else if (tb_cflags(s->base.tb) & CF_PARALLEL) { + gen_helper_stg_parallel(cpu_env, dirty_addr, dirty_addr); + } else { + gen_helper_stg(cpu_env, dirty_addr, dirty_addr); + } + } + + clean_addr = clean_data_tbi(s, dirty_addr); if (is_vector) { if (is_load) { do_fp_ld(s, rt, clean_addr, size); -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> We can simplify our DC_ZVA if we recognize that the largest BS that we actually use in system mode is 64. Let us just assert that it fits within TARGET_PAGE_SIZE. For DC_GVA and STZGM, we want to be able to write whole bytes of tag memory, so assert that BS is >= 2 * TAG_GRANULE, or 32. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-18-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 24 ++++++++++++++++++++++++ 1 file changed, 24 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_realizefn(DeviceState *dev, Error **errp) } #endif + if (tcg_enabled()) { + int dcz_blocklen = 4 << cpu->dcz_blocksize; + + /* + * We only support DCZ blocklen that fits on one page. + * + * Architectually this is always true. However TARGET_PAGE_SIZE + * is variable and, for compatibility with -machine virt-2.7, + * is only 1KiB, as an artifact of legacy ARMv5 subpage support. + * But even then, while the largest architectural DCZ blocklen + * is 2KiB, no cpu actually uses such a large blocklen. + */ + assert(dcz_blocklen <= TARGET_PAGE_SIZE); + + /* + * We only support DCZ blocksize >= 2*TAG_GRANULE, which is to say + * both nibbles of each byte storing tag data may be written at once. + * Since TAG_GRANULE is 16, this means that blocklen must be >= 32. + */ + if (cpu_isar_feature(aa64_mte, cpu)) { + assert(dcz_blocklen >= 2 * TAG_GRANULE); + } + } + qemu_init_vcpu(cs); cpu_reset(cs); -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Now that we know that the operation is on a single page, we need not loop over pages while probing. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-19-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.c | 94 +++++++++++------------------------------ 1 file changed, 25 insertions(+), 69 deletions(-) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -XXX,XX +XXX,XX @@ void HELPER(dc_zva)(CPUARMState *env, uint64_t vaddr_in) * (which matches the usual QEMU behaviour of not implementing either * alignment faults or any memory attribute handling). */ - - ARMCPU *cpu = env_archcpu(env); - uint64_t blocklen = 4 << cpu->dcz_blocksize; + int blocklen = 4 << env_archcpu(env)->dcz_blocksize; uint64_t vaddr = vaddr_in & ~(blocklen - 1); + int mmu_idx = cpu_mmu_index(env, false); + void *mem; + + /* + * Trapless lookup. In addition to actual invalid page, may + * return NULL for I/O, watchpoints, clean pages, etc. + */ + mem = tlb_vaddr_to_host(env, vaddr, MMU_DATA_STORE, mmu_idx); #ifndef CONFIG_USER_ONLY - { + if (unlikely(!mem)) { + uintptr_t ra = GETPC(); + /* - * Slightly awkwardly, QEMU's TARGET_PAGE_SIZE may be less than - * the block size so we might have to do more than one TLB lookup. - * We know that in fact for any v8 CPU the page size is at least 4K - * and the block size must be 2K or less, but TARGET_PAGE_SIZE is only - * 1K as an artefact of legacy v5 subpage support being present in the - * same QEMU executable. So in practice the hostaddr[] array has - * two entries, given the current setting of TARGET_PAGE_BITS_MIN. + * Trap if accessing an invalid page. DC_ZVA requires that we supply + * the original pointer for an invalid page. But watchpoints require + * that we probe the actual space. So do both. */ - int maxidx = DIV_ROUND_UP(blocklen, TARGET_PAGE_SIZE); - void *hostaddr[DIV_ROUND_UP(2 * KiB, 1 << TARGET_PAGE_BITS_MIN)]; - int try, i; - unsigned mmu_idx = cpu_mmu_index(env, false); - TCGMemOpIdx oi = make_memop_idx(MO_UB, mmu_idx); + (void) probe_write(env, vaddr_in, 1, mmu_idx, ra); + mem = probe_write(env, vaddr, blocklen, mmu_idx, ra); - assert(maxidx <= ARRAY_SIZE(hostaddr)); - - for (try = 0; try < 2; try++) { - - for (i = 0; i < maxidx; i++) { - hostaddr[i] = tlb_vaddr_to_host(env, - vaddr + TARGET_PAGE_SIZE * i, - 1, mmu_idx); - if (!hostaddr[i]) { - break; - } - } - if (i == maxidx) { - /* - * If it's all in the TLB it's fair game for just writing to; - * we know we don't need to update dirty status, etc. - */ - for (i = 0; i < maxidx - 1; i++) { - memset(hostaddr[i], 0, TARGET_PAGE_SIZE); - } - memset(hostaddr[i], 0, blocklen - (i * TARGET_PAGE_SIZE)); - return; - } + if (unlikely(!mem)) { /* - * OK, try a store and see if we can populate the tlb. This - * might cause an exception if the memory isn't writable, - * in which case we will longjmp out of here. We must for - * this purpose use the actual register value passed to us - * so that we get the fault address right. + * The only remaining reason for mem == NULL is I/O. + * Just do a series of byte writes as the architecture demands. */ - helper_ret_stb_mmu(env, vaddr_in, 0, oi, GETPC()); - /* Now we can populate the other TLB entries, if any */ - for (i = 0; i < maxidx; i++) { - uint64_t va = vaddr + TARGET_PAGE_SIZE * i; - if (va != (vaddr_in & TARGET_PAGE_MASK)) { - helper_ret_stb_mmu(env, va, 0, oi, GETPC()); - } + for (int i = 0; i < blocklen; i++) { + cpu_stb_mmuidx_ra(env, vaddr + i, 0, mmu_idx, ra); } - } - - /* - * Slow path (probably attempt to do this to an I/O device or - * similar, or clearing of a block of code we have translations - * cached for). Just do a series of byte writes as the architecture - * demands. It's not worth trying to use a cpu_physical_memory_map(), - * memset(), unmap() sequence here because: - * + we'd need to account for the blocksize being larger than a page - * + the direct-RAM access case is almost always going to be dealt - * with in the fastpath code above, so there's no speed benefit - * + we would have to deal with the map returning NULL because the - * bounce buffer was in use - */ - for (i = 0; i < blocklen; i++) { - helper_ret_stb_mmu(env, vaddr + i, 0, oi, GETPC()); + return; } } -#else - memset(g2h(vaddr), 0, blocklen); #endif + + memset(mem, 0, blocklen); } -- 2.20.1 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: 20200626033144.790098-20-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 3 ++ target/arm/translate.h | 2 + target/arm/mte_helper.c | 84 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-a64.c | 72 ++++++++++++++++++++++++++++---- 4 files changed, 153 insertions(+), 8 deletions(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_2(stg_stub, TCG_CALL_NO_WG, void, env, i64) DEF_HELPER_FLAGS_3(st2g, TCG_CALL_NO_WG, void, env, i64, i64) DEF_HELPER_FLAGS_3(st2g_parallel, TCG_CALL_NO_WG, void, env, i64, i64) DEF_HELPER_FLAGS_2(st2g_stub, TCG_CALL_NO_WG, void, env, i64) +DEF_HELPER_FLAGS_2(ldgm, TCG_CALL_NO_WG, i64, env, i64) +DEF_HELPER_FLAGS_3(stgm, TCG_CALL_NO_WG, void, env, i64, i64) +DEF_HELPER_FLAGS_3(stzgm_tags, TCG_CALL_NO_WG, void, env, i64, i64) 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 { * < 0, set by the current instruction. */ int8_t btype; + /* A copy of cpu->dcz_blocksize. */ + uint8_t dcz_blocksize; /* True if this page is guarded. */ bool guarded_page; /* Bottom two bits of XScale c15_cpar coprocessor access control reg */ diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(st2g_stub)(CPUARMState *env, uint64_t ptr) probe_write(env, ptr + TAG_GRANULE, TAG_GRANULE, mmu_idx, ra); } } + +#define LDGM_STGM_SIZE (4 << GMID_EL1_BS) + +uint64_t HELPER(ldgm)(CPUARMState *env, uint64_t ptr) +{ + int mmu_idx = cpu_mmu_index(env, false); + uintptr_t ra = GETPC(); + void *tag_mem; + + ptr = QEMU_ALIGN_DOWN(ptr, LDGM_STGM_SIZE); + + /* Trap if accessing an invalid page. */ + tag_mem = allocation_tag_mem(env, mmu_idx, ptr, MMU_DATA_LOAD, + LDGM_STGM_SIZE, MMU_DATA_LOAD, + LDGM_STGM_SIZE / (2 * TAG_GRANULE), ra); + + /* The tag is squashed to zero if the page does not support tags. */ + if (!tag_mem) { + return 0; + } + + QEMU_BUILD_BUG_ON(GMID_EL1_BS != 6); + /* + * We are loading 64-bits worth of tags. The ordering of elements + * within the word corresponds to a 64-bit little-endian operation. + */ + return ldq_le_p(tag_mem); +} + +void HELPER(stgm)(CPUARMState *env, uint64_t ptr, uint64_t val) +{ + int mmu_idx = cpu_mmu_index(env, false); + uintptr_t ra = GETPC(); + void *tag_mem; + + ptr = QEMU_ALIGN_DOWN(ptr, LDGM_STGM_SIZE); + + /* Trap if accessing an invalid page. */ + tag_mem = allocation_tag_mem(env, mmu_idx, ptr, MMU_DATA_STORE, + LDGM_STGM_SIZE, MMU_DATA_LOAD, + LDGM_STGM_SIZE / (2 * TAG_GRANULE), ra); + + /* + * Tag store only happens if the page support tags, + * and if the OS has enabled access to the tags. + */ + if (!tag_mem) { + return; + } + + QEMU_BUILD_BUG_ON(GMID_EL1_BS != 6); + /* + * We are storing 64-bits worth of tags. The ordering of elements + * within the word corresponds to a 64-bit little-endian operation. + */ + stq_le_p(tag_mem, val); +} + +void HELPER(stzgm_tags)(CPUARMState *env, uint64_t ptr, uint64_t val) +{ + uintptr_t ra = GETPC(); + int mmu_idx = cpu_mmu_index(env, false); + int log2_dcz_bytes, log2_tag_bytes; + intptr_t dcz_bytes, tag_bytes; + uint8_t *mem; + + /* + * In arm_cpu_realizefn, we assert that dcz > LOG2_TAG_GRANULE+1, + * i.e. 32 bytes, which is an unreasonably small dcz anyway, + * to make sure that we can access one complete tag byte here. + */ + log2_dcz_bytes = env_archcpu(env)->dcz_blocksize + 2; + log2_tag_bytes = log2_dcz_bytes - (LOG2_TAG_GRANULE + 1); + dcz_bytes = (intptr_t)1 << log2_dcz_bytes; + tag_bytes = (intptr_t)1 << log2_tag_bytes; + ptr &= -dcz_bytes; + + mem = allocation_tag_mem(env, mmu_idx, ptr, MMU_DATA_STORE, dcz_bytes, + MMU_DATA_STORE, tag_bytes, ra); + if (mem) { + int tag_pair = (val & 0xf) * 0x11; + memset(mem, tag_pair, tag_bytes); + } +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_ldst_tag(DisasContext *s, uint32_t insn) uint64_t offset = sextract64(insn, 12, 9) << LOG2_TAG_GRANULE; int op2 = extract32(insn, 10, 2); int op1 = extract32(insn, 22, 2); - bool is_load = false, is_pair = false, is_zero = false; + bool is_load = false, is_pair = false, is_zero = false, is_mult = false; int index = 0; TCGv_i64 addr, clean_addr, tcg_rt; @@ -XXX,XX +XXX,XX @@ static void disas_ldst_tag(DisasContext *s, uint32_t insn) if (op2 != 0) { /* STG */ index = op2 - 2; - break; + } else { + /* STZGM */ + if (s->current_el == 0 || offset != 0) { + goto do_unallocated; + } + is_mult = is_zero = true; } - goto do_unallocated; + break; case 1: if (op2 != 0) { /* STZG */ @@ -XXX,XX +XXX,XX @@ static void disas_ldst_tag(DisasContext *s, uint32_t insn) /* ST2G */ is_pair = true; index = op2 - 2; - break; + } else { + /* STGM */ + if (s->current_el == 0 || offset != 0) { + goto do_unallocated; + } + is_mult = true; } - goto do_unallocated; + break; case 3: if (op2 != 0) { /* STZ2G */ is_pair = is_zero = true; index = op2 - 2; - break; + } else { + /* LDGM */ + if (s->current_el == 0 || offset != 0) { + goto do_unallocated; + } + is_mult = is_load = true; } - goto do_unallocated; + break; default: do_unallocated: @@ -XXX,XX +XXX,XX @@ static void disas_ldst_tag(DisasContext *s, uint32_t insn) return; } - if (!dc_isar_feature(aa64_mte_insn_reg, s)) { + if (is_mult + ? !dc_isar_feature(aa64_mte, s) + : !dc_isar_feature(aa64_mte_insn_reg, s)) { goto do_unallocated; } @@ -XXX,XX +XXX,XX @@ static void disas_ldst_tag(DisasContext *s, uint32_t insn) tcg_gen_addi_i64(addr, addr, offset); } + if (is_mult) { + tcg_rt = cpu_reg(s, rt); + + if (is_zero) { + int size = 4 << s->dcz_blocksize; + + if (s->ata) { + gen_helper_stzgm_tags(cpu_env, addr, tcg_rt); + } + /* + * The non-tags portion of STZGM is mostly like DC_ZVA, + * except the alignment happens before the access. + */ + clean_addr = clean_data_tbi(s, addr); + tcg_gen_andi_i64(clean_addr, clean_addr, -size); + gen_helper_dc_zva(cpu_env, clean_addr); + } else if (s->ata) { + if (is_load) { + gen_helper_ldgm(tcg_rt, cpu_env, addr); + } else { + gen_helper_stgm(cpu_env, addr, tcg_rt); + } + } else { + MMUAccessType acc = is_load ? MMU_DATA_LOAD : MMU_DATA_STORE; + int size = 4 << GMID_EL1_BS; + + clean_addr = clean_data_tbi(s, addr); + tcg_gen_andi_i64(clean_addr, clean_addr, -size); + gen_probe_access(s, clean_addr, acc, size); + + if (is_load) { + /* The result tags are zeros. */ + tcg_gen_movi_i64(tcg_rt, 0); + } + } + return; + } + if (is_load) { tcg_gen_andi_i64(addr, addr, -TAG_GRANULE); tcg_rt = cpu_reg(s, rt); @@ -XXX,XX +XXX,XX @@ static void aarch64_tr_init_disas_context(DisasContextBase *dcbase, dc->vec_stride = 0; dc->cp_regs = arm_cpu->cp_regs; dc->features = env->features; + dc->dcz_blocksize = arm_cpu->dcz_blocksize; /* Single step state. The code-generation logic here is: * SS_ACTIVE == 0: -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Like the regular data cache flushes, these are nops within qemu. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-21-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 65 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 65 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 @@ static const ARMCPRegInfo mte_reginfo[] = { .opc0 = 3, .opc1 = 3, .crn = 4, .crm = 2, .opc2 = 7, .type = ARM_CP_NO_RAW, .access = PL0_RW, .readfn = tco_read, .writefn = tco_write }, + { .name = "DC_IGVAC", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 6, .opc2 = 3, + .type = ARM_CP_NOP, .access = PL1_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_IGSW", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 6, .opc2 = 4, + .type = ARM_CP_NOP, .access = PL1_W, .accessfn = access_tsw }, + { .name = "DC_IGDVAC", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 6, .opc2 = 5, + .type = ARM_CP_NOP, .access = PL1_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_IGDSW", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 6, .opc2 = 6, + .type = ARM_CP_NOP, .access = PL1_W, .accessfn = access_tsw }, + { .name = "DC_CGSW", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 10, .opc2 = 4, + .type = ARM_CP_NOP, .access = PL1_W, .accessfn = access_tsw }, + { .name = "DC_CGDSW", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 10, .opc2 = 6, + .type = ARM_CP_NOP, .access = PL1_W, .accessfn = access_tsw }, + { .name = "DC_CIGSW", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 14, .opc2 = 4, + .type = ARM_CP_NOP, .access = PL1_W, .accessfn = access_tsw }, + { .name = "DC_CIGDSW", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 14, .opc2 = 6, + .type = ARM_CP_NOP, .access = PL1_W, .accessfn = access_tsw }, REGINFO_SENTINEL }; @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo mte_tco_ro_reginfo[] = { .type = ARM_CP_CONST, .access = PL0_RW, }, REGINFO_SENTINEL }; + +static const ARMCPRegInfo mte_el0_cacheop_reginfo[] = { + { .name = "DC_CGVAC", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 10, .opc2 = 3, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_CGDVAC", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 10, .opc2 = 5, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_CGVAP", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 12, .opc2 = 3, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_CGDVAP", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 12, .opc2 = 5, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_CGVADP", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 13, .opc2 = 3, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_CGDVADP", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 13, .opc2 = 5, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_CIGVAC", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 14, .opc2 = 3, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_CIGDVAC", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 14, .opc2 = 5, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + REGINFO_SENTINEL +}; + #endif static CPAccessResult access_predinv(CPUARMState *env, const ARMCPRegInfo *ri, @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) */ if (cpu_isar_feature(aa64_mte, cpu)) { define_arm_cp_regs(cpu, mte_reginfo); + define_arm_cp_regs(cpu, mte_el0_cacheop_reginfo); } else if (cpu_isar_feature(aa64_mte_insn_reg, cpu)) { define_arm_cp_regs(cpu, mte_tco_ro_reginfo); + define_arm_cp_regs(cpu, mte_el0_cacheop_reginfo); } #endif -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> We will shortly need this in mte_helper.c as well. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-22-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/internals.h | 36 ++++++++++++++++++++++++++++++++++++ target/arm/helper.c | 36 ------------------------------------ 2 files changed, 36 insertions(+), 36 deletions(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ static inline bool regime_is_pan(CPUARMState *env, ARMMMUIdx mmu_idx) } } +/* Return the exception level which controls this address translation regime */ +static inline uint32_t regime_el(CPUARMState *env, ARMMMUIdx mmu_idx) +{ + switch (mmu_idx) { + case ARMMMUIdx_E20_0: + case ARMMMUIdx_E20_2: + case ARMMMUIdx_E20_2_PAN: + case ARMMMUIdx_Stage2: + case ARMMMUIdx_E2: + return 2; + case ARMMMUIdx_SE3: + return 3; + case ARMMMUIdx_SE10_0: + return arm_el_is_aa64(env, 3) ? 1 : 3; + case ARMMMUIdx_SE10_1: + case ARMMMUIdx_SE10_1_PAN: + case ARMMMUIdx_Stage1_E0: + case ARMMMUIdx_Stage1_E1: + case ARMMMUIdx_Stage1_E1_PAN: + case ARMMMUIdx_E10_0: + case ARMMMUIdx_E10_1: + case ARMMMUIdx_E10_1_PAN: + case ARMMMUIdx_MPrivNegPri: + case ARMMMUIdx_MUserNegPri: + case ARMMMUIdx_MPriv: + case ARMMMUIdx_MUser: + case ARMMMUIdx_MSPrivNegPri: + case ARMMMUIdx_MSUserNegPri: + case ARMMMUIdx_MSPriv: + case ARMMMUIdx_MSUser: + return 1; + default: + g_assert_not_reached(); + } +} + /* Return the FSR value for a debug exception (watchpoint, hardware * breakpoint or BKPT insn) targeting the specified exception level. */ 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 arm_cpu_do_interrupt(CPUState *cs) } #endif /* !CONFIG_USER_ONLY */ -/* Return the exception level which controls this address translation regime */ -static uint32_t regime_el(CPUARMState *env, ARMMMUIdx mmu_idx) -{ - switch (mmu_idx) { - case ARMMMUIdx_E20_0: - case ARMMMUIdx_E20_2: - case ARMMMUIdx_E20_2_PAN: - case ARMMMUIdx_Stage2: - case ARMMMUIdx_E2: - return 2; - case ARMMMUIdx_SE3: - return 3; - case ARMMMUIdx_SE10_0: - return arm_el_is_aa64(env, 3) ? 1 : 3; - case ARMMMUIdx_SE10_1: - case ARMMMUIdx_SE10_1_PAN: - case ARMMMUIdx_Stage1_E0: - case ARMMMUIdx_Stage1_E1: - case ARMMMUIdx_Stage1_E1_PAN: - case ARMMMUIdx_E10_0: - case ARMMMUIdx_E10_1: - case ARMMMUIdx_E10_1_PAN: - case ARMMMUIdx_MPrivNegPri: - case ARMMMUIdx_MUserNegPri: - case ARMMMUIdx_MPriv: - case ARMMMUIdx_MUser: - case ARMMMUIdx_MSPrivNegPri: - case ARMMMUIdx_MSUserNegPri: - case ARMMMUIdx_MSPriv: - case ARMMMUIdx_MSUser: - return 1; - default: - g_assert_not_reached(); - } -} - uint64_t arm_sctlr(CPUARMState *env, int el) { /* Only EL0 needs to be adjusted for EL1&0 or EL2&0. */ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> We will shortly need this in mte_helper.c as well. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-23-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/internals.h | 9 +++++++++ target/arm/helper.c | 9 --------- 2 files changed, 9 insertions(+), 9 deletions(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ static inline uint32_t regime_el(CPUARMState *env, ARMMMUIdx mmu_idx) } } +/* Return the TCR controlling this translation regime */ +static inline TCR *regime_tcr(CPUARMState *env, ARMMMUIdx mmu_idx) +{ + if (mmu_idx == ARMMMUIdx_Stage2) { + return &env->cp15.vtcr_el2; + } + return &env->cp15.tcr_el[regime_el(env, mmu_idx)]; +} + /* Return the FSR value for a debug exception (watchpoint, hardware * breakpoint or BKPT insn) targeting the specified exception level. */ 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 inline uint64_t regime_ttbr(CPUARMState *env, ARMMMUIdx mmu_idx, #endif /* !CONFIG_USER_ONLY */ -/* Return the TCR controlling this translation regime */ -static inline TCR *regime_tcr(CPUARMState *env, ARMMMUIdx mmu_idx) -{ - if (mmu_idx == ARMMMUIdx_Stage2) { - return &env->cp15.vtcr_el2; - } - return &env->cp15.tcr_el[regime_el(env, mmu_idx)]; -} - /* Convert a possible stage1+2 MMU index into the appropriate * stage 1 MMU index */ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Replace existing uses of check_data_tbi in translate-a64.c that perform a single logical memory access. Leave the helper blank for now to reduce the patch size. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-24-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 1 + target/arm/internals.h | 8 +++ target/arm/translate-a64.h | 2 + target/arm/mte_helper.c | 8 +++ target/arm/translate-a64.c | 100 ++++++++++++++++++++++++++++--------- 5 files changed, 95 insertions(+), 24 deletions(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(autdb, TCG_CALL_NO_WG, i64, env, i64, i64) DEF_HELPER_FLAGS_2(xpaci, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_2(xpacd, TCG_CALL_NO_RWG_SE, i64, env, i64) +DEF_HELPER_FLAGS_3(mte_check1, TCG_CALL_NO_WG, i64, env, i32, i64) DEF_HELPER_FLAGS_3(irg, TCG_CALL_NO_RWG, i64, env, i64, i64) DEF_HELPER_FLAGS_4(addsubg, TCG_CALL_NO_RWG_SE, i64, env, i64, s32, i32) DEF_HELPER_FLAGS_3(ldg, TCG_CALL_NO_WG, i64, env, i64, i64) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ void arm_log_exception(int idx); #define LOG2_TAG_GRANULE 4 #define TAG_GRANULE (1 << LOG2_TAG_GRANULE) +/* Bits within a descriptor passed to the helper_mte_check* functions. */ +FIELD(MTEDESC, MIDX, 0, 4) +FIELD(MTEDESC, TBI, 4, 2) +FIELD(MTEDESC, TCMA, 6, 2) +FIELD(MTEDESC, WRITE, 8, 1) +FIELD(MTEDESC, ESIZE, 9, 5) +FIELD(MTEDESC, TSIZE, 14, 10) /* mte_checkN only */ + static inline int allocation_tag_from_addr(uint64_t ptr) { return extract64(ptr, 56, 4); 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 @@ TCGv_ptr get_fpstatus_ptr(bool); bool logic_imm_decode_wmask(uint64_t *result, unsigned int immn, unsigned int imms, unsigned int immr); bool sve_access_check(DisasContext *s); +TCGv_i64 gen_mte_check1(DisasContext *s, TCGv_i64 addr, bool is_write, + bool tag_checked, int log2_size); /* We should have at some point before trying to access an FP register * done the necessary access check, so assert that diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(stzgm_tags)(CPUARMState *env, uint64_t ptr, uint64_t val) memset(mem, tag_pair, tag_bytes); } } + +/* + * Perform an MTE checked access for a single logical or atomic access. + */ +uint64_t HELPER(mte_check1)(CPUARMState *env, uint32_t desc, uint64_t ptr) +{ + return ptr; +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void gen_a64_set_pc(DisasContext *s, TCGv_i64 src) } /* - * Return a "clean" address for ADDR according to TBID. - * This is always a fresh temporary, as we need to be able to - * increment this independently of a dirty write-back address. + * Handle MTE and/or TBI. + * + * For TBI, ideally, we would do nothing. Proper behaviour on fault is + * for the tag to be present in the FAR_ELx register. But for user-only + * mode we do not have a TLB with which to implement this, so we must + * remove the top byte now. + * + * Always return a fresh temporary that we can increment independently + * of the write-back address. */ + static TCGv_i64 clean_data_tbi(DisasContext *s, TCGv_i64 addr) { TCGv_i64 clean = new_tmp_a64(s); - /* - * In order to get the correct value in the FAR_ELx register, - * we must present the memory subsystem with the "dirty" address - * including the TBI. In system mode we can make this work via - * the TLB, dropping the TBI during translation. But for user-only - * mode we don't have that option, and must remove the top byte now. - */ #ifdef CONFIG_USER_ONLY gen_top_byte_ignore(s, clean, addr, s->tbid); #else @@ -XXX,XX +XXX,XX @@ static void gen_probe_access(DisasContext *s, TCGv_i64 ptr, tcg_temp_free_i32(t_size); } +/* + * For MTE, check a single logical or atomic access. This probes a single + * address, the exact one specified. The size and alignment of the access + * is not relevant to MTE, per se, but watchpoints do require the size, + * and we want to recognize those before making any other changes to state. + */ +static TCGv_i64 gen_mte_check1_mmuidx(DisasContext *s, TCGv_i64 addr, + bool is_write, bool tag_checked, + int log2_size, bool is_unpriv, + int core_idx) +{ + if (tag_checked && s->mte_active[is_unpriv]) { + TCGv_i32 tcg_desc; + TCGv_i64 ret; + int desc = 0; + + desc = FIELD_DP32(desc, MTEDESC, MIDX, core_idx); + desc = FIELD_DP32(desc, MTEDESC, TBI, s->tbid); + desc = FIELD_DP32(desc, MTEDESC, TCMA, s->tcma); + desc = FIELD_DP32(desc, MTEDESC, WRITE, is_write); + desc = FIELD_DP32(desc, MTEDESC, ESIZE, 1 << log2_size); + tcg_desc = tcg_const_i32(desc); + + ret = new_tmp_a64(s); + gen_helper_mte_check1(ret, cpu_env, tcg_desc, addr); + tcg_temp_free_i32(tcg_desc); + + return ret; + } + return clean_data_tbi(s, addr); +} + +TCGv_i64 gen_mte_check1(DisasContext *s, TCGv_i64 addr, bool is_write, + bool tag_checked, int log2_size) +{ + return gen_mte_check1_mmuidx(s, addr, is_write, tag_checked, log2_size, + false, get_mem_index(s)); +} + typedef struct DisasCompare64 { TCGCond cond; TCGv_i64 value; @@ -XXX,XX +XXX,XX @@ static void gen_compare_and_swap(DisasContext *s, int rs, int rt, if (rn == 31) { gen_check_sp_alignment(s); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), true, rn != 31, size); tcg_gen_atomic_cmpxchg_i64(tcg_rs, clean_addr, tcg_rs, tcg_rt, memidx, size | MO_ALIGN | s->be_data); } @@ -XXX,XX +XXX,XX @@ static void gen_compare_and_swap_pair(DisasContext *s, int rs, int rt, if (rn == 31) { gen_check_sp_alignment(s); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + + /* This is a single atomic access, despite the "pair". */ + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), true, rn != 31, size + 1); if (size == 2) { TCGv_i64 cmp = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) if (is_lasr) { tcg_gen_mb(TCG_MO_ALL | TCG_BAR_STRL); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), + true, rn != 31, size); gen_store_exclusive(s, rs, rt, rt2, clean_addr, size, false); return; @@ -XXX,XX +XXX,XX @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) if (rn == 31) { gen_check_sp_alignment(s); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), + false, rn != 31, size); s->is_ldex = true; gen_load_exclusive(s, rt, rt2, clean_addr, size, false); if (is_lasr) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) gen_check_sp_alignment(s); } tcg_gen_mb(TCG_MO_ALL | TCG_BAR_STRL); - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), + true, rn != 31, size); do_gpr_st(s, cpu_reg(s, rt), clean_addr, size, true, rt, disas_ldst_compute_iss_sf(size, false, 0), is_lasr); return; @@ -XXX,XX +XXX,XX @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) if (rn == 31) { gen_check_sp_alignment(s); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), + false, rn != 31, size); do_gpr_ld(s, cpu_reg(s, rt), clean_addr, size, false, false, true, rt, disas_ldst_compute_iss_sf(size, false, 0), is_lasr); tcg_gen_mb(TCG_MO_ALL | TCG_BAR_LDAQ); @@ -XXX,XX +XXX,XX @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) if (is_lasr) { tcg_gen_mb(TCG_MO_ALL | TCG_BAR_STRL); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), + true, rn != 31, size); gen_store_exclusive(s, rs, rt, rt2, clean_addr, size, true); return; } @@ -XXX,XX +XXX,XX @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) if (rn == 31) { gen_check_sp_alignment(s); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), + false, rn != 31, size); s->is_ldex = true; gen_load_exclusive(s, rt, rt2, clean_addr, size, true); if (is_lasr) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_reg_imm9(DisasContext *s, uint32_t insn, bool iss_valid = !is_vector; bool post_index; bool writeback; + int memidx; TCGv_i64 clean_addr, dirty_addr; @@ -XXX,XX +XXX,XX @@ static void disas_ldst_reg_imm9(DisasContext *s, uint32_t insn, if (!post_index) { tcg_gen_addi_i64(dirty_addr, dirty_addr, imm9); } - clean_addr = clean_data_tbi(s, dirty_addr); + + memidx = is_unpriv ? get_a64_user_mem_index(s) : get_mem_index(s); + clean_addr = gen_mte_check1_mmuidx(s, dirty_addr, is_store, + writeback || rn != 31, + size, is_unpriv, memidx); if (is_vector) { if (is_store) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_reg_imm9(DisasContext *s, uint32_t insn, } } else { TCGv_i64 tcg_rt = cpu_reg(s, rt); - int memidx = is_unpriv ? get_a64_user_mem_index(s) : get_mem_index(s); bool iss_sf = disas_ldst_compute_iss_sf(size, is_signed, opc); if (is_store) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_reg_roffset(DisasContext *s, uint32_t insn, ext_and_shift_reg(tcg_rm, tcg_rm, opt, shift ? size : 0); tcg_gen_add_i64(dirty_addr, dirty_addr, tcg_rm); - clean_addr = clean_data_tbi(s, dirty_addr); + clean_addr = gen_mte_check1(s, dirty_addr, is_store, true, size); if (is_vector) { if (is_store) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_reg_unsigned_imm(DisasContext *s, uint32_t insn, dirty_addr = read_cpu_reg_sp(s, rn, 1); offset = imm12 << size; tcg_gen_addi_i64(dirty_addr, dirty_addr, offset); - clean_addr = clean_data_tbi(s, dirty_addr); + clean_addr = gen_mte_check1(s, dirty_addr, is_store, rn != 31, size); if (is_vector) { if (is_store) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_atomic(DisasContext *s, uint32_t insn, if (rn == 31) { gen_check_sp_alignment(s); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), false, rn != 31, size); if (o3_opc == 014) { /* @@ -XXX,XX +XXX,XX @@ static void disas_ldst_pac(DisasContext *s, uint32_t insn, tcg_gen_addi_i64(dirty_addr, dirty_addr, offset); /* Note that "clean" and "dirty" here refer to TBI not PAC. */ - clean_addr = clean_data_tbi(s, dirty_addr); + clean_addr = gen_mte_check1(s, dirty_addr, false, + is_wback || rn != 31, size); tcg_rt = cpu_reg(s, rt); do_gpr_ld(s, tcg_rt, clean_addr, size, /* is_signed */ false, -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Replace existing uses of check_data_tbi in translate-a64.c that perform multiple logical memory access. Leave the helper blank for now to reduce the patch size. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-25-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 1 + target/arm/translate-a64.h | 2 ++ target/arm/mte_helper.c | 8 +++++ target/arm/translate-a64.c | 71 +++++++++++++++++++++++++++++--------- 4 files changed, 66 insertions(+), 16 deletions(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_2(xpaci, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_2(xpacd, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_3(mte_check1, TCG_CALL_NO_WG, i64, env, i32, i64) +DEF_HELPER_FLAGS_3(mte_checkN, TCG_CALL_NO_WG, i64, env, i32, i64) DEF_HELPER_FLAGS_3(irg, TCG_CALL_NO_RWG, i64, env, i64, i64) DEF_HELPER_FLAGS_4(addsubg, TCG_CALL_NO_RWG_SE, i64, env, i64, s32, i32) DEF_HELPER_FLAGS_3(ldg, TCG_CALL_NO_WG, i64, env, i64, i64) 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 @@ bool logic_imm_decode_wmask(uint64_t *result, unsigned int immn, bool sve_access_check(DisasContext *s); TCGv_i64 gen_mte_check1(DisasContext *s, TCGv_i64 addr, bool is_write, bool tag_checked, int log2_size); +TCGv_i64 gen_mte_checkN(DisasContext *s, TCGv_i64 addr, bool is_write, + bool tag_checked, int count, int log2_esize); /* We should have at some point before trying to access an FP register * done the necessary access check, so assert that diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(mte_check1)(CPUARMState *env, uint32_t desc, uint64_t ptr) { return ptr; } + +/* + * Perform an MTE checked access for multiple logical accesses. + */ +uint64_t HELPER(mte_checkN)(CPUARMState *env, uint32_t desc, uint64_t ptr) +{ + return ptr; +} 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 @@ TCGv_i64 gen_mte_check1(DisasContext *s, TCGv_i64 addr, bool is_write, false, get_mem_index(s)); } +/* + * For MTE, check multiple logical sequential accesses. + */ +TCGv_i64 gen_mte_checkN(DisasContext *s, TCGv_i64 addr, bool is_write, + bool tag_checked, int log2_esize, int total_size) +{ + if (tag_checked && s->mte_active[0] && total_size != (1 << log2_esize)) { + TCGv_i32 tcg_desc; + TCGv_i64 ret; + int desc = 0; + + 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, is_write); + desc = FIELD_DP32(desc, MTEDESC, ESIZE, 1 << log2_esize); + desc = FIELD_DP32(desc, MTEDESC, TSIZE, total_size); + tcg_desc = tcg_const_i32(desc); + + ret = new_tmp_a64(s); + gen_helper_mte_checkN(ret, cpu_env, tcg_desc, addr); + tcg_temp_free_i32(tcg_desc); + + return ret; + } + return gen_mte_check1(s, addr, is_write, tag_checked, log2_esize); +} + typedef struct DisasCompare64 { TCGCond cond; TCGv_i64 value; @@ -XXX,XX +XXX,XX @@ static void disas_ldst_pair(DisasContext *s, uint32_t insn) } } - clean_addr = clean_data_tbi(s, dirty_addr); + clean_addr = gen_mte_checkN(s, dirty_addr, !is_load, + (wback || rn != 31) && !set_tag, + size, 2 << size); + if (is_vector) { if (is_load) { do_fp_ld(s, rt, clean_addr, size); @@ -XXX,XX +XXX,XX @@ static void disas_ldst_multiple_struct(DisasContext *s, uint32_t insn) TCGv_i64 clean_addr, tcg_rn, tcg_ebytes; MemOp endian = s->be_data; - int ebytes; /* bytes per element */ + int total; /* total bytes */ int elements; /* elements per vector */ int rpt; /* num iterations */ int selem; /* structure elements */ @@ -XXX,XX +XXX,XX @@ static void disas_ldst_multiple_struct(DisasContext *s, uint32_t insn) endian = MO_LE; } - /* Consecutive little-endian elements from a single register + total = rpt * selem * (is_q ? 16 : 8); + tcg_rn = cpu_reg_sp(s, rn); + + /* + * Issue the MTE check vs the logical repeat count, before we + * promote consecutive little-endian elements below. + */ + clean_addr = gen_mte_checkN(s, tcg_rn, is_store, is_postidx || rn != 31, + size, total); + + /* + * Consecutive little-endian elements from a single register * can be promoted to a larger little-endian operation. */ if (selem == 1 && endian == MO_LE) { size = 3; } - ebytes = 1 << size; - elements = (is_q ? 16 : 8) / ebytes; - - tcg_rn = cpu_reg_sp(s, rn); - clean_addr = clean_data_tbi(s, tcg_rn); - tcg_ebytes = tcg_const_i64(ebytes); + elements = (is_q ? 16 : 8) >> size; + tcg_ebytes = tcg_const_i64(1 << size); for (r = 0; r < rpt; r++) { int e; for (e = 0; e < elements; e++) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_multiple_struct(DisasContext *s, uint32_t insn) if (is_postidx) { if (rm == 31) { - tcg_gen_addi_i64(tcg_rn, tcg_rn, rpt * elements * selem * ebytes); + tcg_gen_addi_i64(tcg_rn, tcg_rn, total); } else { tcg_gen_add_i64(tcg_rn, tcg_rn, cpu_reg(s, rm)); } @@ -XXX,XX +XXX,XX @@ static void disas_ldst_single_struct(DisasContext *s, uint32_t insn) int selem = (extract32(opc, 0, 1) << 1 | R) + 1; bool replicate = false; int index = is_q << 3 | S << 2 | size; - int ebytes, xs; + int xs, total; TCGv_i64 clean_addr, tcg_rn, tcg_ebytes; if (extract32(insn, 31, 1)) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_single_struct(DisasContext *s, uint32_t insn) return; } - ebytes = 1 << scale; - if (rn == 31) { gen_check_sp_alignment(s); } + total = selem << scale; tcg_rn = cpu_reg_sp(s, rn); - clean_addr = clean_data_tbi(s, tcg_rn); - tcg_ebytes = tcg_const_i64(ebytes); + clean_addr = gen_mte_checkN(s, tcg_rn, !is_load, is_postidx || rn != 31, + scale, total); + + tcg_ebytes = tcg_const_i64(1 << scale); for (xs = 0; xs < selem; xs++) { if (replicate) { /* Load and replicate to all elements */ @@ -XXX,XX +XXX,XX @@ static void disas_ldst_single_struct(DisasContext *s, uint32_t insn) if (is_postidx) { if (rm == 31) { - tcg_gen_addi_i64(tcg_rn, tcg_rn, selem * ebytes); + tcg_gen_addi_i64(tcg_rn, tcg_rn, total); } else { tcg_gen_add_i64(tcg_rn, tcg_rn, cpu_reg(s, rm)); } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Fill out the stub that was added earlier. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-26-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/internals.h | 48 +++++++++++++++ target/arm/mte_helper.c | 132 +++++++++++++++++++++++++++++++++++++++- 2 files changed, 179 insertions(+), 1 deletion(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ FIELD(MTEDESC, WRITE, 8, 1) FIELD(MTEDESC, ESIZE, 9, 5) FIELD(MTEDESC, TSIZE, 14, 10) /* mte_checkN only */ +bool mte_probe1(CPUARMState *env, uint32_t desc, uint64_t ptr); +uint64_t mte_check1(CPUARMState *env, uint32_t desc, + uint64_t ptr, uintptr_t ra); + static inline int allocation_tag_from_addr(uint64_t ptr) { return extract64(ptr, 56, 4); @@ -XXX,XX +XXX,XX @@ static inline uint64_t address_with_allocation_tag(uint64_t ptr, int rtag) return deposit64(ptr, 56, 4, rtag); } +/* Return true if tbi bits mean that the access is checked. */ +static inline bool tbi_check(uint32_t desc, int bit55) +{ + return (desc >> (R_MTEDESC_TBI_SHIFT + bit55)) & 1; +} + +/* Return true if tcma bits mean that the access is unchecked. */ +static inline bool tcma_check(uint32_t desc, int bit55, int ptr_tag) +{ + /* + * We had extracted bit55 and ptr_tag for other reasons, so fold + * (ptr<59:55> == 00000 || ptr<59:55> == 11111) into a single test. + */ + bool match = ((ptr_tag + bit55) & 0xf) == 0; + bool tcma = (desc >> (R_MTEDESC_TCMA_SHIFT + bit55)) & 1; + return tcma && match; +} + +/* + * For TBI, ideally, we would do nothing. Proper behaviour on fault is + * for the tag to be present in the FAR_ELx register. But for user-only + * mode, we do not have a TLB with which to implement this, so we must + * remove the top byte. + */ +static inline uint64_t useronly_clean_ptr(uint64_t ptr) +{ + /* TBI is known to be enabled. */ +#ifdef CONFIG_USER_ONLY + ptr = sextract64(ptr, 0, 56); +#endif + return ptr; +} + +static inline uint64_t useronly_maybe_clean_ptr(uint32_t desc, uint64_t ptr) +{ +#ifdef CONFIG_USER_ONLY + int64_t clean_ptr = sextract64(ptr, 0, 56); + if (tbi_check(desc, clean_ptr < 0)) { + ptr = clean_ptr; + } +#endif + return ptr; +} + #endif diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(stzgm_tags)(CPUARMState *env, uint64_t ptr, uint64_t val) } } +/* Record a tag check failure. */ +static void mte_check_fail(CPUARMState *env, int mmu_idx, + uint64_t dirty_ptr, uintptr_t ra) +{ + ARMMMUIdx arm_mmu_idx = core_to_aa64_mmu_idx(mmu_idx); + int el, reg_el, tcf, select; + uint64_t sctlr; + + reg_el = regime_el(env, arm_mmu_idx); + sctlr = env->cp15.sctlr_el[reg_el]; + + switch (arm_mmu_idx) { + case ARMMMUIdx_E10_0: + case ARMMMUIdx_E20_0: + el = 0; + tcf = extract64(sctlr, 38, 2); + break; + default: + el = reg_el; + tcf = extract64(sctlr, 40, 2); + } + + switch (tcf) { + case 1: + /* + * Tag check fail causes a synchronous exception. + * + * In restore_state_to_opc, we set the exception syndrome + * for the load or store operation. Unwind first so we + * may overwrite that with the syndrome for the tag check. + */ + cpu_restore_state(env_cpu(env), ra, true); + env->exception.vaddress = dirty_ptr; + raise_exception(env, EXCP_DATA_ABORT, + syn_data_abort_no_iss(el != 0, 0, 0, 0, 0, 0, 0x11), + exception_target_el(env)); + /* noreturn, but fall through to the assert anyway */ + + case 0: + /* + * Tag check fail does not affect the PE. + * We eliminate this case by not setting MTE_ACTIVE + * in tb_flags, so that we never make this runtime call. + */ + g_assert_not_reached(); + + case 2: + /* Tag check fail causes asynchronous flag set. */ + mmu_idx = arm_mmu_idx_el(env, el); + if (regime_has_2_ranges(mmu_idx)) { + select = extract64(dirty_ptr, 55, 1); + } else { + select = 0; + } + env->cp15.tfsr_el[el] |= 1 << select; + break; + + default: + /* Case 3: Reserved. */ + qemu_log_mask(LOG_GUEST_ERROR, + "Tag check failure with SCTLR_EL%d.TCF%s " + "set to reserved value %d\n", + reg_el, el ? "" : "0", tcf); + break; + } +} + /* * Perform an MTE checked access for a single logical or atomic access. */ +static bool mte_probe1_int(CPUARMState *env, uint32_t desc, uint64_t ptr, + uintptr_t ra, int bit55) +{ + int mem_tag, mmu_idx, ptr_tag, size; + MMUAccessType type; + uint8_t *mem; + + ptr_tag = allocation_tag_from_addr(ptr); + + if (tcma_check(desc, bit55, ptr_tag)) { + return true; + } + + mmu_idx = FIELD_EX32(desc, MTEDESC, MIDX); + type = FIELD_EX32(desc, MTEDESC, WRITE) ? MMU_DATA_STORE : MMU_DATA_LOAD; + size = FIELD_EX32(desc, MTEDESC, ESIZE); + + mem = allocation_tag_mem(env, mmu_idx, ptr, type, size, + MMU_DATA_LOAD, 1, ra); + if (!mem) { + return true; + } + + mem_tag = load_tag1(ptr, mem); + return ptr_tag == mem_tag; +} + +/* + * No-fault version of mte_check1, to be used by SVE for MemSingleNF. + * Returns false if the access is Checked and the check failed. This + * is only intended to probe the tag -- the validity of the page must + * be checked beforehand. + */ +bool mte_probe1(CPUARMState *env, uint32_t desc, uint64_t ptr) +{ + int bit55 = extract64(ptr, 55, 1); + + /* If TBI is disabled, the access is unchecked. */ + if (unlikely(!tbi_check(desc, bit55))) { + return true; + } + + return mte_probe1_int(env, desc, ptr, 0, bit55); +} + +uint64_t mte_check1(CPUARMState *env, uint32_t desc, + uint64_t ptr, uintptr_t ra) +{ + int bit55 = extract64(ptr, 55, 1); + + /* If TBI is disabled, the access is unchecked, and ptr is not dirty. */ + if (unlikely(!tbi_check(desc, bit55))) { + return ptr; + } + + if (unlikely(!mte_probe1_int(env, desc, ptr, ra, bit55))) { + int mmu_idx = FIELD_EX32(desc, MTEDESC, MIDX); + mte_check_fail(env, mmu_idx, ptr, ra); + } + + return useronly_clean_ptr(ptr); +} + uint64_t HELPER(mte_check1)(CPUARMState *env, uint32_t desc, uint64_t ptr) { - return ptr; + return mte_check1(env, desc, ptr, GETPC()); } /* -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Fill out the stub that was added earlier. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-27-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/internals.h | 2 + target/arm/mte_helper.c | 165 +++++++++++++++++++++++++++++++++++++++- 2 files changed, 166 insertions(+), 1 deletion(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ FIELD(MTEDESC, TSIZE, 14, 10) /* mte_checkN only */ bool mte_probe1(CPUARMState *env, uint32_t desc, uint64_t ptr); uint64_t mte_check1(CPUARMState *env, uint32_t desc, uint64_t ptr, uintptr_t ra); +uint64_t mte_checkN(CPUARMState *env, uint32_t desc, + uint64_t ptr, uintptr_t ra); static inline int allocation_tag_from_addr(uint64_t ptr) { diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(mte_check1)(CPUARMState *env, uint32_t desc, uint64_t ptr) /* * Perform an MTE checked access for multiple logical accesses. */ + +/** + * checkN: + * @tag: tag memory to test + * @odd: true to begin testing at tags at odd nibble + * @cmp: the tag to compare against + * @count: number of tags to test + * + * Return the number of successful tests. + * Thus a return value < @count indicates a failure. + * + * A note about sizes: count is expected to be small. + * + * The most common use will be LDP/STP of two integer registers, + * which means 16 bytes of memory touching at most 2 tags, but + * often the access is aligned and thus just 1 tag. + * + * Using AdvSIMD LD/ST (multiple), one can access 64 bytes of memory, + * touching at most 5 tags. SVE LDR/STR (vector) with the default + * vector length is also 64 bytes; the maximum architectural length + * is 256 bytes touching at most 9 tags. + * + * The loop below uses 7 logical operations and 1 memory operation + * per tag pair. An implementation that loads an aligned word and + * uses masking to ignore adjacent tags requires 18 logical operations + * and thus does not begin to pay off until 6 tags. + * Which, according to the survey above, is unlikely to be common. + */ +static int checkN(uint8_t *mem, int odd, int cmp, int count) +{ + int n = 0, diff; + + /* Replicate the test tag and compare. */ + cmp *= 0x11; + diff = *mem++ ^ cmp; + + if (odd) { + goto start_odd; + } + + while (1) { + /* Test even tag. */ + if (unlikely((diff) & 0x0f)) { + break; + } + if (++n == count) { + break; + } + + start_odd: + /* Test odd tag. */ + if (unlikely((diff) & 0xf0)) { + break; + } + if (++n == count) { + break; + } + + diff = *mem++ ^ cmp; + } + return n; +} + +uint64_t mte_checkN(CPUARMState *env, uint32_t desc, + uint64_t ptr, uintptr_t ra) +{ + int mmu_idx, ptr_tag, bit55; + uint64_t ptr_last, ptr_end, prev_page, next_page; + uint64_t tag_first, tag_end; + uint64_t tag_byte_first, tag_byte_end; + uint32_t esize, total, tag_count, tag_size, n, c; + uint8_t *mem1, *mem2; + MMUAccessType type; + + bit55 = extract64(ptr, 55, 1); + + /* If TBI is disabled, the access is unchecked, and ptr is not dirty. */ + if (unlikely(!tbi_check(desc, bit55))) { + return ptr; + } + + ptr_tag = allocation_tag_from_addr(ptr); + + if (tcma_check(desc, bit55, ptr_tag)) { + goto done; + } + + mmu_idx = FIELD_EX32(desc, MTEDESC, MIDX); + type = FIELD_EX32(desc, MTEDESC, WRITE) ? MMU_DATA_STORE : MMU_DATA_LOAD; + esize = FIELD_EX32(desc, MTEDESC, ESIZE); + total = FIELD_EX32(desc, MTEDESC, TSIZE); + + /* Find the addr of the end of the access, and of the last element. */ + ptr_end = ptr + total; + ptr_last = ptr_end - esize; + + /* Round the bounds to the tag granule, and compute the number of tags. */ + tag_first = QEMU_ALIGN_DOWN(ptr, TAG_GRANULE); + tag_end = QEMU_ALIGN_UP(ptr_last, TAG_GRANULE); + tag_count = (tag_end - tag_first) / TAG_GRANULE; + + /* Round the bounds to twice the tag granule, and compute the bytes. */ + tag_byte_first = QEMU_ALIGN_DOWN(ptr, 2 * TAG_GRANULE); + tag_byte_end = QEMU_ALIGN_UP(ptr_last, 2 * TAG_GRANULE); + + /* Locate the page boundaries. */ + prev_page = ptr & TARGET_PAGE_MASK; + next_page = prev_page + TARGET_PAGE_SIZE; + + if (likely(tag_end - prev_page <= TARGET_PAGE_SIZE)) { + /* Memory access stays on one page. */ + tag_size = (tag_byte_end - tag_byte_first) / (2 * TAG_GRANULE); + mem1 = allocation_tag_mem(env, mmu_idx, ptr, type, total, + MMU_DATA_LOAD, tag_size, ra); + if (!mem1) { + goto done; + } + /* Perform all of the comparisons. */ + n = checkN(mem1, ptr & TAG_GRANULE, ptr_tag, tag_count); + } else { + /* Memory access crosses to next page. */ + tag_size = (next_page - tag_byte_first) / (2 * TAG_GRANULE); + mem1 = allocation_tag_mem(env, mmu_idx, ptr, type, next_page - ptr, + MMU_DATA_LOAD, tag_size, ra); + + tag_size = (tag_byte_end - next_page) / (2 * TAG_GRANULE); + mem2 = allocation_tag_mem(env, mmu_idx, next_page, type, + ptr_end - next_page, + MMU_DATA_LOAD, tag_size, ra); + + /* + * Perform all of the comparisons. + * Note the possible but unlikely case of the operation spanning + * two pages that do not both have tagging enabled. + */ + n = c = (next_page - tag_first) / TAG_GRANULE; + if (mem1) { + n = checkN(mem1, ptr & TAG_GRANULE, ptr_tag, c); + } + if (n == c) { + if (!mem2) { + goto done; + } + n += checkN(mem2, 0, ptr_tag, tag_count - c); + } + } + + /* + * If we failed, we know which granule. Compute the element that + * is first in that granule, and signal failure on that element. + */ + if (unlikely(n < tag_count)) { + uint64_t fail_ofs; + + fail_ofs = tag_first + n * TAG_GRANULE - ptr; + fail_ofs = ROUND_UP(fail_ofs, esize); + mte_check_fail(env, mmu_idx, ptr + fail_ofs, ra); + } + + done: + return useronly_clean_ptr(ptr); +} + uint64_t HELPER(mte_checkN)(CPUARMState *env, uint32_t desc, uint64_t ptr) { - return ptr; + return mte_checkN(env, desc, ptr, GETPC()); } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Use a special helper for DC_ZVA, rather than the more general mte_checkN. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-28-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 1 + target/arm/mte_helper.c | 106 +++++++++++++++++++++++++++++++++++++ target/arm/translate-a64.c | 16 +++++- 3 files changed, 122 insertions(+), 1 deletion(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_2(xpacd, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_3(mte_check1, TCG_CALL_NO_WG, i64, env, i32, i64) DEF_HELPER_FLAGS_3(mte_checkN, TCG_CALL_NO_WG, i64, env, i32, i64) +DEF_HELPER_FLAGS_3(mte_check_zva, TCG_CALL_NO_WG, i64, env, i32, i64) DEF_HELPER_FLAGS_3(irg, TCG_CALL_NO_RWG, i64, env, i64, i64) DEF_HELPER_FLAGS_4(addsubg, TCG_CALL_NO_RWG_SE, i64, env, i64, s32, i32) DEF_HELPER_FLAGS_3(ldg, TCG_CALL_NO_WG, i64, env, i64, i64) diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(mte_checkN)(CPUARMState *env, uint32_t desc, uint64_t ptr) { return mte_checkN(env, desc, ptr, GETPC()); } + +/* + * Perform an MTE checked access for DC_ZVA. + */ +uint64_t HELPER(mte_check_zva)(CPUARMState *env, uint32_t desc, uint64_t ptr) +{ + uintptr_t ra = GETPC(); + int log2_dcz_bytes, log2_tag_bytes; + int mmu_idx, bit55; + intptr_t dcz_bytes, tag_bytes, i; + void *mem; + uint64_t ptr_tag, mem_tag, align_ptr; + + bit55 = extract64(ptr, 55, 1); + + /* If TBI is disabled, the access is unchecked, and ptr is not dirty. */ + if (unlikely(!tbi_check(desc, bit55))) { + return ptr; + } + + ptr_tag = allocation_tag_from_addr(ptr); + + if (tcma_check(desc, bit55, ptr_tag)) { + goto done; + } + + /* + * In arm_cpu_realizefn, we asserted that dcz > LOG2_TAG_GRANULE+1, + * i.e. 32 bytes, which is an unreasonably small dcz anyway, to make + * sure that we can access one complete tag byte here. + */ + log2_dcz_bytes = env_archcpu(env)->dcz_blocksize + 2; + log2_tag_bytes = log2_dcz_bytes - (LOG2_TAG_GRANULE + 1); + dcz_bytes = (intptr_t)1 << log2_dcz_bytes; + tag_bytes = (intptr_t)1 << log2_tag_bytes; + align_ptr = ptr & -dcz_bytes; + + /* + * Trap if accessing an invalid page. DC_ZVA requires that we supply + * the original pointer for an invalid page. But watchpoints require + * that we probe the actual space. So do both. + */ + mmu_idx = FIELD_EX32(desc, MTEDESC, MIDX); + (void) probe_write(env, ptr, 1, mmu_idx, ra); + mem = allocation_tag_mem(env, mmu_idx, align_ptr, MMU_DATA_STORE, + dcz_bytes, MMU_DATA_LOAD, tag_bytes, ra); + if (!mem) { + goto done; + } + + /* + * Unlike the reasoning for checkN, DC_ZVA is always aligned, and thus + * it is quite easy to perform all of the comparisons at once without + * any extra masking. + * + * The most common zva block size is 64; some of the thunderx cpus use + * a block size of 128. For user-only, aarch64_max_initfn will set the + * block size to 512. Fill out the other cases for future-proofing. + * + * In order to be able to find the first miscompare later, we want the + * tag bytes to be in little-endian order. + */ + switch (log2_tag_bytes) { + case 0: /* zva_blocksize 32 */ + mem_tag = *(uint8_t *)mem; + ptr_tag *= 0x11u; + break; + case 1: /* zva_blocksize 64 */ + mem_tag = cpu_to_le16(*(uint16_t *)mem); + ptr_tag *= 0x1111u; + break; + case 2: /* zva_blocksize 128 */ + mem_tag = cpu_to_le32(*(uint32_t *)mem); + ptr_tag *= 0x11111111u; + break; + case 3: /* zva_blocksize 256 */ + mem_tag = cpu_to_le64(*(uint64_t *)mem); + ptr_tag *= 0x1111111111111111ull; + break; + + default: /* zva_blocksize 512, 1024, 2048 */ + ptr_tag *= 0x1111111111111111ull; + i = 0; + do { + mem_tag = cpu_to_le64(*(uint64_t *)(mem + i)); + if (unlikely(mem_tag != ptr_tag)) { + goto fail; + } + i += 8; + align_ptr += 16 * TAG_GRANULE; + } while (i < tag_bytes); + goto done; + } + + if (likely(mem_tag == ptr_tag)) { + goto done; + } + + fail: + /* Locate the first nibble that differs. */ + i = ctz64(mem_tag ^ ptr_tag) >> 4; + mte_check_fail(env, mmu_idx, align_ptr + i * TAG_GRANULE, ra); + + done: + return useronly_clean_ptr(ptr); +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread, return; case ARM_CP_DC_ZVA: /* Writes clear the aligned block of memory which rt points into. */ - tcg_rt = clean_data_tbi(s, cpu_reg(s, rt)); + if (s->mte_active[0]) { + TCGv_i32 t_desc; + int desc = 0; + + 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); + t_desc = tcg_const_i32(desc); + + tcg_rt = new_tmp_a64(s); + gen_helper_mte_check_zva(tcg_rt, cpu_env, t_desc, cpu_reg(s, rt)); + tcg_temp_free_i32(t_desc); + } else { + tcg_rt = clean_data_tbi(s, cpu_reg(s, rt)); + } gen_helper_dc_zva(cpu_env, tcg_rt); return; default: -- 2.20.1 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: 20200626033144.790098-29-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 61 +++++++++++++++++++++----------------- 1 file changed, 33 insertions(+), 28 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) int len_remain = len % 8; int nparts = len / 8 + ctpop8(len_remain); int midx = get_mem_index(s); - TCGv_i64 addr, t0, t1; + TCGv_i64 dirty_addr, clean_addr, t0, t1; - addr = tcg_temp_new_i64(); - t0 = tcg_temp_new_i64(); + dirty_addr = tcg_temp_new_i64(); + tcg_gen_addi_i64(dirty_addr, cpu_reg_sp(s, rn), imm); + clean_addr = gen_mte_checkN(s, dirty_addr, false, rn != 31, len, MO_8); + tcg_temp_free_i64(dirty_addr); - /* Note that unpredicated load/store of vector/predicate registers + /* + * Note that unpredicated load/store of vector/predicate registers * are defined as a stream of bytes, which equates to little-endian - * operations on larger quantities. There is no nice way to force - * a little-endian load for aarch64_be-linux-user out of line. - * + * operations on larger quantities. * Attempt to keep code expansion to a minimum by limiting the * amount of unrolling done. */ if (nparts <= 4) { int i; + t0 = tcg_temp_new_i64(); for (i = 0; i < len_align; i += 8) { - tcg_gen_addi_i64(addr, cpu_reg_sp(s, rn), imm + i); - tcg_gen_qemu_ld_i64(t0, addr, midx, MO_LEQ); + tcg_gen_qemu_ld_i64(t0, clean_addr, midx, MO_LEQ); tcg_gen_st_i64(t0, cpu_env, vofs + i); + tcg_gen_addi_i64(clean_addr, cpu_reg_sp(s, rn), 8); } + tcg_temp_free_i64(t0); } else { TCGLabel *loop = gen_new_label(); TCGv_ptr tp, i = tcg_const_local_ptr(0); + /* Copy the clean address into a local temp, live across the loop. */ + t0 = clean_addr; + clean_addr = tcg_temp_local_new_i64(); + tcg_gen_mov_i64(clean_addr, t0); + tcg_temp_free_i64(t0); + gen_set_label(loop); - /* Minimize the number of local temps that must be re-read from - * the stack each iteration. Instead, re-compute values other - * than the loop counter. - */ + t0 = tcg_temp_new_i64(); + tcg_gen_qemu_ld_i64(t0, clean_addr, midx, MO_LEQ); + tcg_gen_addi_i64(clean_addr, clean_addr, 8); + tp = tcg_temp_new_ptr(); - tcg_gen_addi_ptr(tp, i, imm); - tcg_gen_extu_ptr_i64(addr, tp); - tcg_gen_add_i64(addr, addr, cpu_reg_sp(s, rn)); - - tcg_gen_qemu_ld_i64(t0, addr, midx, MO_LEQ); - tcg_gen_add_ptr(tp, cpu_env, i); tcg_gen_addi_ptr(i, i, 8); tcg_gen_st_i64(t0, tp, vofs); tcg_temp_free_ptr(tp); + tcg_temp_free_i64(t0); tcg_gen_brcondi_ptr(TCG_COND_LTU, i, len_align, loop); tcg_temp_free_ptr(i); } - /* Predicate register loads can be any multiple of 2. + /* + * Predicate register loads can be any multiple of 2. * Note that we still store the entire 64-bit unit into cpu_env. */ if (len_remain) { - tcg_gen_addi_i64(addr, cpu_reg_sp(s, rn), imm + len_align); - + t0 = tcg_temp_new_i64(); switch (len_remain) { case 2: case 4: case 8: - tcg_gen_qemu_ld_i64(t0, addr, midx, MO_LE | ctz32(len_remain)); + tcg_gen_qemu_ld_i64(t0, clean_addr, midx, + MO_LE | ctz32(len_remain)); break; case 6: t1 = tcg_temp_new_i64(); - tcg_gen_qemu_ld_i64(t0, addr, midx, MO_LEUL); - tcg_gen_addi_i64(addr, addr, 4); - tcg_gen_qemu_ld_i64(t1, addr, midx, MO_LEUW); + tcg_gen_qemu_ld_i64(t0, clean_addr, midx, MO_LEUL); + tcg_gen_addi_i64(clean_addr, clean_addr, 4); + tcg_gen_qemu_ld_i64(t1, clean_addr, midx, MO_LEUW); tcg_gen_deposit_i64(t0, t0, t1, 32, 32); tcg_temp_free_i64(t1); break; @@ -XXX,XX +XXX,XX @@ static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) g_assert_not_reached(); } tcg_gen_st_i64(t0, cpu_env, vofs + len_align); + tcg_temp_free_i64(t0); } - tcg_temp_free_i64(addr); - tcg_temp_free_i64(t0); + tcg_temp_free_i64(clean_addr); } /* Similarly for stores. */ -- 2.20.1 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: 20200626033144.790098-30-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 61 +++++++++++++++++++++----------------- 1 file changed, 33 insertions(+), 28 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static void do_str(DisasContext *s, uint32_t vofs, int len, int rn, int imm) int len_remain = len % 8; int nparts = len / 8 + ctpop8(len_remain); int midx = get_mem_index(s); - TCGv_i64 addr, t0; + TCGv_i64 dirty_addr, clean_addr, t0; - addr = tcg_temp_new_i64(); - t0 = tcg_temp_new_i64(); + dirty_addr = tcg_temp_new_i64(); + tcg_gen_addi_i64(dirty_addr, cpu_reg_sp(s, rn), imm); + clean_addr = gen_mte_checkN(s, dirty_addr, false, rn != 31, len, MO_8); + tcg_temp_free_i64(dirty_addr); /* Note that unpredicated load/store of vector/predicate registers * are defined as a stream of bytes, which equates to little-endian @@ -XXX,XX +XXX,XX @@ static void do_str(DisasContext *s, uint32_t vofs, int len, int rn, int imm) if (nparts <= 4) { int i; + t0 = tcg_temp_new_i64(); for (i = 0; i < len_align; i += 8) { tcg_gen_ld_i64(t0, cpu_env, vofs + i); - tcg_gen_addi_i64(addr, cpu_reg_sp(s, rn), imm + i); - tcg_gen_qemu_st_i64(t0, addr, midx, MO_LEQ); + tcg_gen_qemu_st_i64(t0, clean_addr, midx, MO_LEQ); + tcg_gen_addi_i64(clean_addr, cpu_reg_sp(s, rn), 8); } + tcg_temp_free_i64(t0); } else { TCGLabel *loop = gen_new_label(); - TCGv_ptr t2, i = tcg_const_local_ptr(0); + TCGv_ptr tp, i = tcg_const_local_ptr(0); + + /* Copy the clean address into a local temp, live across the loop. */ + t0 = clean_addr; + clean_addr = tcg_temp_local_new_i64(); + tcg_gen_mov_i64(clean_addr, t0); + tcg_temp_free_i64(t0); gen_set_label(loop); - t2 = tcg_temp_new_ptr(); - tcg_gen_add_ptr(t2, cpu_env, i); - tcg_gen_ld_i64(t0, t2, vofs); - - /* Minimize the number of local temps that must be re-read from - * the stack each iteration. Instead, re-compute values other - * than the loop counter. - */ - tcg_gen_addi_ptr(t2, i, imm); - tcg_gen_extu_ptr_i64(addr, t2); - tcg_gen_add_i64(addr, addr, cpu_reg_sp(s, rn)); - tcg_temp_free_ptr(t2); - - tcg_gen_qemu_st_i64(t0, addr, midx, MO_LEQ); - + t0 = tcg_temp_new_i64(); + tp = tcg_temp_new_ptr(); + tcg_gen_add_ptr(tp, cpu_env, i); + tcg_gen_ld_i64(t0, tp, vofs); tcg_gen_addi_ptr(i, i, 8); + tcg_temp_free_ptr(tp); + + tcg_gen_qemu_st_i64(t0, clean_addr, midx, MO_LEQ); + tcg_gen_addi_i64(clean_addr, clean_addr, 8); + tcg_temp_free_i64(t0); tcg_gen_brcondi_ptr(TCG_COND_LTU, i, len_align, loop); tcg_temp_free_ptr(i); @@ -XXX,XX +XXX,XX @@ static void do_str(DisasContext *s, uint32_t vofs, int len, int rn, int imm) /* 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_addi_i64(addr, cpu_reg_sp(s, rn), imm + len_align); switch (len_remain) { case 2: case 4: case 8: - tcg_gen_qemu_st_i64(t0, addr, midx, MO_LE | ctz32(len_remain)); + tcg_gen_qemu_st_i64(t0, clean_addr, midx, + MO_LE | ctz32(len_remain)); break; case 6: - tcg_gen_qemu_st_i64(t0, addr, midx, MO_LEUL); - tcg_gen_addi_i64(addr, addr, 4); + tcg_gen_qemu_st_i64(t0, clean_addr, midx, MO_LEUL); + tcg_gen_addi_i64(clean_addr, clean_addr, 4); tcg_gen_shri_i64(t0, t0, 32); - tcg_gen_qemu_st_i64(t0, addr, midx, MO_LEUW); + tcg_gen_qemu_st_i64(t0, clean_addr, midx, MO_LEUW); break; default: g_assert_not_reached(); } + tcg_temp_free_i64(t0); } - tcg_temp_free_i64(addr); - tcg_temp_free_i64(t0); + tcg_temp_free_i64(clean_addr); } static bool trans_LDR_zri(DisasContext *s, arg_rri *a) -- 2.20.1 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: 20200626033144.790098-31-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_LD1R_zpri(DisasContext *s, arg_rpri_load *a) unsigned esz = dtype_esz[a->dtype]; unsigned msz = dtype_msz(a->dtype); TCGLabel *over = gen_new_label(); - TCGv_i64 temp; + TCGv_i64 temp, clean_addr; /* If the guarding predicate has no bits set, no load occurs. */ if (psz <= 8) { @@ -XXX,XX +XXX,XX @@ static bool trans_LD1R_zpri(DisasContext *s, arg_rpri_load *a) /* Load the data. */ temp = tcg_temp_new_i64(); tcg_gen_addi_i64(temp, cpu_reg_sp(s, a->rn), a->imm << msz); - tcg_gen_qemu_ld_i64(temp, temp, get_mem_index(s), + clean_addr = gen_mte_check1(s, temp, false, true, msz); + + tcg_gen_qemu_ld_i64(temp, clean_addr, get_mem_index(s), s->be_data | dtype_mop[a->dtype]); /* Broadcast to *all* elements. */ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Move the variable declarations to the top of the function, but do not create a new label before sve_access_check. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-32-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_LD1RQ_zpri(DisasContext *s, arg_rpri_load *a) /* Load and broadcast element. */ static bool trans_LD1R_zpri(DisasContext *s, arg_rpri_load *a) { - if (!sve_access_check(s)) { - return true; - } - unsigned vsz = vec_full_reg_size(s); unsigned psz = pred_full_reg_size(s); unsigned esz = dtype_esz[a->dtype]; unsigned msz = dtype_msz(a->dtype); - TCGLabel *over = gen_new_label(); + TCGLabel *over; TCGv_i64 temp, clean_addr; + if (!sve_access_check(s)) { + return true; + } + + over = gen_new_label(); + /* If the guarding predicate has no bits set, no load occurs. */ if (psz <= 8) { /* Reduce the pred_esz_masks value simply to reduce the -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Introduce an lvalue macro to wrap target_tlb_bit0. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-33-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 13 +++++++++++++ target/arm/helper.c | 2 +- target/arm/translate-a64.c | 2 +- 3 files changed, 15 insertions(+), 2 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 uint64_t *aa64_vfp_qreg(CPUARMState *env, unsigned regno) /* Shared between translate-sve.c and sve_helper.c. */ extern const uint64_t pred_esz_masks[4]; +/* Helper for the macros below, validating the argument type. */ +static inline MemTxAttrs *typecheck_memtxattrs(MemTxAttrs *x) +{ + return x; +} + +/* + * Lvalue macros for ARM TLB bits that we must cache in the TCG TLB. + * Using these should be a bit more self-documenting than using the + * generic target bits directly. + */ +#define arm_tlb_bti_gp(x) (typecheck_memtxattrs(x)->target_tlb_bit0) + /* * Naming convention for isar_feature functions: * Functions which test 32-bit ID registers should have _aa32_ in diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static bool get_phys_addr_lpae(CPUARMState *env, target_ulong address, } /* When in aarch64 mode, and BTI is enabled, remember GP in the IOTLB. */ if (aarch64 && guarded && cpu_isar_feature(aa64_bti, cpu)) { - txattrs->target_tlb_bit0 = true; + arm_tlb_bti_gp(txattrs) = true; } if (cacheattrs != NULL) { 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 is_guarded_page(CPUARMState *env, DisasContext *s) * table entry even for that case. */ return (tlb_hit(entry->addr_code, addr) && - env_tlb(env)->d[mmu_idx].iotlb[index].attrs.target_tlb_bit0); + arm_tlb_bti_gp(&env_tlb(env)->d[mmu_idx].iotlb[index].attrs)); #endif } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Because the elements are sequential, we can eliminate many tests all at once when the tag hits TCMA, or if the page(s) are not Tagged. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-34-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 1 + target/arm/helper-sve.h | 58 ++++++++++ target/arm/internals.h | 6 + target/arm/sve_helper.c | 218 ++++++++++++++++++++++++++++++------- target/arm/translate-sve.c | 186 ++++++++++++++++++++++--------- 5 files changed, 378 insertions(+), 91 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 MemTxAttrs *typecheck_memtxattrs(MemTxAttrs *x) * generic target bits directly. */ #define arm_tlb_bti_gp(x) (typecheck_memtxattrs(x)->target_tlb_bit0) +#define arm_tlb_mte_tagged(x) (typecheck_memtxattrs(x)->target_tlb_bit1) /* * Naming convention for isar_feature functions: 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_ld1sds_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ld1sdu_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ld1sds_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1bhu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1bsu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1bdu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1bhs_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1bss_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1bds_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1hsu_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hdu_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hds_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1hsu_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hdu_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hds_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1sdu_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1sds_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1sdu_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1sds_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + DEF_HELPER_FLAGS_4(sve_ldff1bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldff1bhu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldff1bsu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ void arm_log_exception(int idx); #define LOG2_TAG_GRANULE 4 #define TAG_GRANULE (1 << LOG2_TAG_GRANULE) +/* + * The SVE simd_data field, for memory ops, contains either + * rd (5 bits) or a shift count (2 bits). + */ +#define SVE_MTEDESC_SHIFT 5 + /* Bits within a descriptor passed to the helper_mte_check* functions. */ FIELD(MTEDESC, MIDX, 0, 4) FIELD(MTEDESC, TBI, 4, 2) diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ static void sve_cont_ldst_watchpoints(SVEContLdSt *info, CPUARMState *env, #endif } +typedef uint64_t mte_check_fn(CPUARMState *, uint32_t, uint64_t, uintptr_t); + +static inline QEMU_ALWAYS_INLINE +void sve_cont_ldst_mte_check_int(SVEContLdSt *info, CPUARMState *env, + uint64_t *vg, target_ulong addr, int esize, + int msize, uint32_t mtedesc, uintptr_t ra, + mte_check_fn *check) +{ + intptr_t mem_off, reg_off, reg_last; + + /* Process the page only if MemAttr == Tagged. */ + if (arm_tlb_mte_tagged(&info->page[0].attrs)) { + mem_off = info->mem_off_first[0]; + reg_off = info->reg_off_first[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) { + check(env, mtedesc, addr, ra); + } + reg_off += esize; + mem_off += msize; + } while (reg_off <= reg_last && (reg_off & 63)); + } while (reg_off <= reg_last); + } + + mem_off = info->mem_off_first[1]; + if (mem_off >= 0 && arm_tlb_mte_tagged(&info->page[1].attrs)) { + reg_off = info->reg_off_first[1]; + reg_last = info->reg_off_last[1]; + + do { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + check(env, mtedesc, addr, ra); + } + reg_off += esize; + mem_off += msize; + } while (reg_off & 63); + } while (reg_off <= reg_last); + } +} + +typedef void sve_cont_ldst_mte_check_fn(SVEContLdSt *info, CPUARMState *env, + uint64_t *vg, target_ulong addr, + int esize, int msize, uint32_t mtedesc, + uintptr_t ra); + +static void sve_cont_ldst_mte_check1(SVEContLdSt *info, CPUARMState *env, + uint64_t *vg, target_ulong addr, + int esize, int msize, uint32_t mtedesc, + uintptr_t ra) +{ + sve_cont_ldst_mte_check_int(info, env, vg, addr, esize, msize, + mtedesc, ra, mte_check1); +} + +static void sve_cont_ldst_mte_checkN(SVEContLdSt *info, CPUARMState *env, + uint64_t *vg, target_ulong addr, + int esize, int msize, uint32_t mtedesc, + uintptr_t ra) +{ + sve_cont_ldst_mte_check_int(info, env, vg, addr, esize, msize, + mtedesc, ra, mte_checkN); +} + + /* * Common helper for all contiguous 1,2,3,4-register predicated stores. */ static inline QEMU_ALWAYS_INLINE void sve_ldN_r(CPUARMState *env, uint64_t *vg, const target_ulong addr, uint32_t desc, const uintptr_t retaddr, - const int esz, const int msz, const int N, + const int esz, const int msz, const int N, uint32_t mtedesc, sve_ldst1_host_fn *host_fn, - sve_ldst1_tlb_fn *tlb_fn) + sve_ldst1_tlb_fn *tlb_fn, + sve_cont_ldst_mte_check_fn *mte_check_fn) { const unsigned rd = simd_data(desc); const intptr_t reg_max = simd_oprsz(desc); @@ -XXX,XX +XXX,XX @@ void sve_ldN_r(CPUARMState *env, uint64_t *vg, const target_ulong addr, sve_cont_ldst_watchpoints(&info, env, vg, addr, 1 << esz, N << msz, BP_MEM_READ, retaddr); - /* TODO: MTE check. */ + /* + * Handle mte checks for all active elements. + * Since TBI must be set for MTE, !mtedesc => !mte_active. + */ + if (mte_check_fn && mtedesc) { + mte_check_fn(&info, env, vg, addr, 1 << esz, N << msz, + mtedesc, retaddr); + } flags = info.page[0].flags | info.page[1].flags; if (unlikely(flags != 0)) { @@ -XXX,XX +XXX,XX @@ void sve_ldN_r(CPUARMState *env, uint64_t *vg, const target_ulong addr, } } -#define DO_LD1_1(NAME, ESZ) \ -void HELPER(sve_##NAME##_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, MO_8, 1, \ - sve_##NAME##_host, sve_##NAME##_tlb); \ +static inline QEMU_ALWAYS_INLINE +void sve_ldN_r_mte(CPUARMState *env, uint64_t *vg, target_ulong addr, + uint32_t desc, const uintptr_t ra, + const int esz, const int msz, const int N, + 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; + } + + sve_ldN_r(env, vg, addr, desc, ra, esz, msz, N, mtedesc, host_fn, tlb_fn, + N == 1 ? sve_cont_ldst_mte_check1 : sve_cont_ldst_mte_checkN); } -#define DO_LD1_2(NAME, ESZ, MSZ) \ -void HELPER(sve_##NAME##_le_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, 1, \ - sve_##NAME##_le_host, sve_##NAME##_le_tlb); \ -} \ -void HELPER(sve_##NAME##_be_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, 1, \ - sve_##NAME##_be_host, sve_##NAME##_be_tlb); \ +#define DO_LD1_1(NAME, ESZ) \ +void HELPER(sve_##NAME##_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, MO_8, 1, 0, \ + sve_##NAME##_host, sve_##NAME##_tlb, NULL); \ +} \ +void HELPER(sve_##NAME##_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r_mte(env, vg, addr, desc, GETPC(), ESZ, MO_8, 1, \ + sve_##NAME##_host, sve_##NAME##_tlb); \ +} + +#define DO_LD1_2(NAME, ESZ, MSZ) \ +void HELPER(sve_##NAME##_le_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, 1, 0, \ + sve_##NAME##_le_host, sve_##NAME##_le_tlb, NULL); \ +} \ +void HELPER(sve_##NAME##_be_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, 1, 0, \ + sve_##NAME##_be_host, sve_##NAME##_be_tlb, NULL); \ +} \ +void HELPER(sve_##NAME##_le_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, 1, \ + sve_##NAME##_le_host, sve_##NAME##_le_tlb); \ +} \ +void HELPER(sve_##NAME##_be_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, 1, \ + sve_##NAME##_be_host, sve_##NAME##_be_tlb); \ } DO_LD1_1(ld1bb, MO_8) @@ -XXX,XX +XXX,XX @@ DO_LD1_2(ld1dd, MO_64, MO_64) #undef DO_LD1_1 #undef DO_LD1_2 -#define DO_LDN_1(N) \ -void HELPER(sve_ld##N##bb_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_ldN_r(env, vg, addr, desc, GETPC(), MO_8, MO_8, N, \ - sve_ld1bb_host, sve_ld1bb_tlb); \ +#define DO_LDN_1(N) \ +void HELPER(sve_ld##N##bb_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r(env, vg, addr, desc, GETPC(), MO_8, MO_8, N, 0, \ + sve_ld1bb_host, sve_ld1bb_tlb, NULL); \ +} \ +void HELPER(sve_ld##N##bb_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r_mte(env, vg, addr, desc, GETPC(), MO_8, MO_8, N, \ + sve_ld1bb_host, sve_ld1bb_tlb); \ } -#define DO_LDN_2(N, SUFF, ESZ) \ -void HELPER(sve_ld##N##SUFF##_le_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, ESZ, N, \ - sve_ld1##SUFF##_le_host, sve_ld1##SUFF##_le_tlb); \ -} \ -void HELPER(sve_ld##N##SUFF##_be_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, ESZ, N, \ - sve_ld1##SUFF##_be_host, sve_ld1##SUFF##_be_tlb); \ +#define DO_LDN_2(N, SUFF, ESZ) \ +void HELPER(sve_ld##N##SUFF##_le_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, ESZ, N, 0, \ + sve_ld1##SUFF##_le_host, sve_ld1##SUFF##_le_tlb, NULL); \ +} \ +void HELPER(sve_ld##N##SUFF##_be_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, ESZ, N, 0, \ + sve_ld1##SUFF##_be_host, sve_ld1##SUFF##_be_tlb, NULL); \ +} \ +void HELPER(sve_ld##N##SUFF##_le_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r_mte(env, vg, addr, desc, GETPC(), ESZ, ESZ, N, \ + sve_ld1##SUFF##_le_host, sve_ld1##SUFF##_le_tlb); \ +} \ +void HELPER(sve_ld##N##SUFF##_be_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r_mte(env, vg, addr, desc, GETPC(), ESZ, ESZ, N, \ + sve_ld1##SUFF##_be_host, sve_ld1##SUFF##_be_tlb); \ } DO_LDN_1(2) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static const uint8_t dtype_esz[16] = { }; static void do_mem_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, - int dtype, gen_helper_gvec_mem *fn) + int dtype, uint32_t mte_n, bool is_write, + gen_helper_gvec_mem *fn) { unsigned vsz = vec_full_reg_size(s); TCGv_ptr t_pg; TCGv_i32 t_desc; - int desc; + int desc = 0; - /* For e.g. LD4, there are not enough arguments to pass all 4 + /* + * For e.g. LD4, there are not enough arguments to pass all 4 * registers as pointers, so encode the regno into the data field. * For consistency, do this even for LD1. + * TODO: mte_n check here while callers are updated. */ - desc = simd_desc(vsz, vsz, zt); + if (mte_n && s->mte_active[0]) { + int msz = dtype_msz(dtype); + + 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, is_write); + desc = FIELD_DP32(desc, MTEDESC, ESIZE, 1 << msz); + desc = FIELD_DP32(desc, MTEDESC, TSIZE, mte_n << msz); + desc <<= SVE_MTEDESC_SHIFT; + } + desc = simd_desc(vsz, vsz, zt | desc); t_desc = tcg_const_i32(desc); t_pg = tcg_temp_new_ptr(); @@ -XXX,XX +XXX,XX @@ static void do_mem_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, static void do_ld_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, int dtype, int nreg) { - static gen_helper_gvec_mem * const fns[2][16][4] = { - /* Little-endian */ - { { gen_helper_sve_ld1bb_r, gen_helper_sve_ld2bb_r, + static gen_helper_gvec_mem * const fns[2][2][16][4] = { + { /* mte inactive, little-endian */ + { { gen_helper_sve_ld1bb_r, gen_helper_sve_ld2bb_r, gen_helper_sve_ld3bb_r, gen_helper_sve_ld4bb_r }, - { gen_helper_sve_ld1bhu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bsu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bdu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bhu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bsu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bdu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1sds_le_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hh_le_r, gen_helper_sve_ld2hh_le_r, - gen_helper_sve_ld3hh_le_r, gen_helper_sve_ld4hh_le_r }, - { gen_helper_sve_ld1hsu_le_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hdu_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1sds_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hh_le_r, gen_helper_sve_ld2hh_le_r, + gen_helper_sve_ld3hh_le_r, gen_helper_sve_ld4hh_le_r }, + { gen_helper_sve_ld1hsu_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hdu_le_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hds_le_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hss_le_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1ss_le_r, gen_helper_sve_ld2ss_le_r, - gen_helper_sve_ld3ss_le_r, gen_helper_sve_ld4ss_le_r }, - { gen_helper_sve_ld1sdu_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hds_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hss_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1ss_le_r, gen_helper_sve_ld2ss_le_r, + gen_helper_sve_ld3ss_le_r, gen_helper_sve_ld4ss_le_r }, + { gen_helper_sve_ld1sdu_le_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bds_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bss_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bhs_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1dd_le_r, gen_helper_sve_ld2dd_le_r, - gen_helper_sve_ld3dd_le_r, gen_helper_sve_ld4dd_le_r } }, + { gen_helper_sve_ld1bds_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bss_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bhs_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1dd_le_r, gen_helper_sve_ld2dd_le_r, + gen_helper_sve_ld3dd_le_r, gen_helper_sve_ld4dd_le_r } }, - /* Big-endian */ - { { gen_helper_sve_ld1bb_r, gen_helper_sve_ld2bb_r, - gen_helper_sve_ld3bb_r, gen_helper_sve_ld4bb_r }, - { gen_helper_sve_ld1bhu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bsu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bdu_r, NULL, NULL, NULL }, + /* mte inactive, big-endian */ + { { gen_helper_sve_ld1bb_r, gen_helper_sve_ld2bb_r, + gen_helper_sve_ld3bb_r, gen_helper_sve_ld4bb_r }, + { gen_helper_sve_ld1bhu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bsu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bdu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1sds_be_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hh_be_r, gen_helper_sve_ld2hh_be_r, - gen_helper_sve_ld3hh_be_r, gen_helper_sve_ld4hh_be_r }, - { gen_helper_sve_ld1hsu_be_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hdu_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1sds_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hh_be_r, gen_helper_sve_ld2hh_be_r, + gen_helper_sve_ld3hh_be_r, gen_helper_sve_ld4hh_be_r }, + { gen_helper_sve_ld1hsu_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hdu_be_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hds_be_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hss_be_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1ss_be_r, gen_helper_sve_ld2ss_be_r, - gen_helper_sve_ld3ss_be_r, gen_helper_sve_ld4ss_be_r }, - { gen_helper_sve_ld1sdu_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hds_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hss_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1ss_be_r, gen_helper_sve_ld2ss_be_r, + gen_helper_sve_ld3ss_be_r, gen_helper_sve_ld4ss_be_r }, + { gen_helper_sve_ld1sdu_be_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bds_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bss_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bhs_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1dd_be_r, gen_helper_sve_ld2dd_be_r, - gen_helper_sve_ld3dd_be_r, gen_helper_sve_ld4dd_be_r } } + { gen_helper_sve_ld1bds_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bss_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bhs_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1dd_be_r, gen_helper_sve_ld2dd_be_r, + gen_helper_sve_ld3dd_be_r, gen_helper_sve_ld4dd_be_r } } }, + + { /* mte active, little-endian */ + { { gen_helper_sve_ld1bb_r_mte, + gen_helper_sve_ld2bb_r_mte, + gen_helper_sve_ld3bb_r_mte, + gen_helper_sve_ld4bb_r_mte }, + { gen_helper_sve_ld1bhu_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bsu_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bdu_r_mte, NULL, NULL, NULL }, + + { gen_helper_sve_ld1sds_le_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1hh_le_r_mte, + gen_helper_sve_ld2hh_le_r_mte, + gen_helper_sve_ld3hh_le_r_mte, + gen_helper_sve_ld4hh_le_r_mte }, + { gen_helper_sve_ld1hsu_le_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1hdu_le_r_mte, NULL, NULL, NULL }, + + { gen_helper_sve_ld1hds_le_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1hss_le_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1ss_le_r_mte, + gen_helper_sve_ld2ss_le_r_mte, + gen_helper_sve_ld3ss_le_r_mte, + gen_helper_sve_ld4ss_le_r_mte }, + { gen_helper_sve_ld1sdu_le_r_mte, NULL, NULL, NULL }, + + { gen_helper_sve_ld1bds_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bss_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bhs_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1dd_le_r_mte, + gen_helper_sve_ld2dd_le_r_mte, + gen_helper_sve_ld3dd_le_r_mte, + gen_helper_sve_ld4dd_le_r_mte } }, + + /* mte active, big-endian */ + { { gen_helper_sve_ld1bb_r_mte, + gen_helper_sve_ld2bb_r_mte, + gen_helper_sve_ld3bb_r_mte, + gen_helper_sve_ld4bb_r_mte }, + { gen_helper_sve_ld1bhu_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bsu_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bdu_r_mte, NULL, NULL, NULL }, + + { gen_helper_sve_ld1sds_be_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1hh_be_r_mte, + gen_helper_sve_ld2hh_be_r_mte, + gen_helper_sve_ld3hh_be_r_mte, + gen_helper_sve_ld4hh_be_r_mte }, + { gen_helper_sve_ld1hsu_be_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1hdu_be_r_mte, NULL, NULL, NULL }, + + { gen_helper_sve_ld1hds_be_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1hss_be_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1ss_be_r_mte, + gen_helper_sve_ld2ss_be_r_mte, + gen_helper_sve_ld3ss_be_r_mte, + gen_helper_sve_ld4ss_be_r_mte }, + { gen_helper_sve_ld1sdu_be_r_mte, NULL, NULL, NULL }, + + { gen_helper_sve_ld1bds_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bss_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bhs_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1dd_be_r_mte, + gen_helper_sve_ld2dd_be_r_mte, + gen_helper_sve_ld3dd_be_r_mte, + gen_helper_sve_ld4dd_be_r_mte } } }, }; - gen_helper_gvec_mem *fn = fns[s->be_data == MO_BE][dtype][nreg]; + gen_helper_gvec_mem *fn + = fns[s->mte_active[0]][s->be_data == MO_BE][dtype][nreg]; - /* While there are holes in the table, they are not + /* + * While there are holes in the table, they are not * accessible via the instruction encoding. */ assert(fn != NULL); - do_mem_zpa(s, zt, pg, addr, dtype, fn); + do_mem_zpa(s, zt, pg, addr, dtype, nreg, false, fn); } static bool trans_LD_zprr(DisasContext *s, arg_rprr_load *a) @@ -XXX,XX +XXX,XX @@ static bool trans_LDFF1_zprr(DisasContext *s, arg_rprr_load *a) TCGv_i64 addr = new_tmp_a64(s); tcg_gen_shli_i64(addr, cpu_reg(s, a->rm), dtype_msz(a->dtype)); tcg_gen_add_i64(addr, addr, cpu_reg_sp(s, a->rn)); - do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, + do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, 0, false, fns[s->be_data == MO_BE][a->dtype]); } return true; @@ -XXX,XX +XXX,XX @@ static bool trans_LDNF1_zpri(DisasContext *s, arg_rpri_load *a) TCGv_i64 addr = new_tmp_a64(s); tcg_gen_addi_i64(addr, cpu_reg_sp(s, a->rn), off); - do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, + do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, 0, false, fns[s->be_data == MO_BE][a->dtype]); } return true; @@ -XXX,XX +XXX,XX @@ static void do_st_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, fn = fn_multiple[be][nreg - 1][msz]; } assert(fn != NULL); - do_mem_zpa(s, zt, pg, addr, msz_dtype(s, msz), fn); + do_mem_zpa(s, zt, pg, addr, msz_dtype(s, msz), 0, true, fn); } static bool trans_ST_zprr(DisasContext *s, arg_rprr_store *a) -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Because the elements are sequential, we can eliminate many tests all at once when the tag hits TCMA, or if the page(s) are not Tagged. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-35-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sve.h | 47 +++++++++++ target/arm/sve_helper.c | 95 ++++++++++++++++------ target/arm/translate-sve.c | 162 ++++++++++++++++++++++++------------- 3 files changed, 226 insertions(+), 78 deletions(-) diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sve_st1hd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st1sd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st1sd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1bh_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1bs_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1bd_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1hs_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1hd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1hs_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1hd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1sd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1sd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + DEF_HELPER_FLAGS_6(sve_ldbsu_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldhsu_le_zsu, TCG_CALL_NO_WG, diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ DO_LDFF1_LDNF1_2(dd, MO_64, MO_64) */ static inline QEMU_ALWAYS_INLINE -void sve_stN_r(CPUARMState *env, uint64_t *vg, target_ulong addr, uint32_t desc, - const uintptr_t retaddr, const int esz, - const int msz, const int N, +void sve_stN_r(CPUARMState *env, uint64_t *vg, target_ulong addr, + uint32_t desc, const uintptr_t retaddr, + const int esz, const int msz, const int N, uint32_t mtedesc, sve_ldst1_host_fn *host_fn, - sve_ldst1_tlb_fn *tlb_fn) + sve_ldst1_tlb_fn *tlb_fn, + sve_cont_ldst_mte_check_fn *mte_check_fn) { const unsigned rd = simd_data(desc); const intptr_t reg_max = simd_oprsz(desc); @@ -XXX,XX +XXX,XX @@ void sve_stN_r(CPUARMState *env, uint64_t *vg, target_ulong addr, uint32_t desc, sve_cont_ldst_watchpoints(&info, env, vg, addr, 1 << esz, N << msz, BP_MEM_WRITE, retaddr); - /* TODO: MTE check. */ + /* + * Handle mte checks for all active elements. + * Since TBI must be set for MTE, !mtedesc => !mte_active. + */ + if (mte_check_fn && mtedesc) { + mte_check_fn(&info, env, vg, addr, 1 << esz, N << msz, + mtedesc, retaddr); + } flags = info.page[0].flags | info.page[1].flags; if (unlikely(flags != 0)) { @@ -XXX,XX +XXX,XX @@ void sve_stN_r(CPUARMState *env, uint64_t *vg, target_ulong addr, uint32_t desc, } } -#define DO_STN_1(N, NAME, ESZ) \ -void HELPER(sve_st##N##NAME##_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_stN_r(env, vg, addr, desc, GETPC(), ESZ, MO_8, N, \ - sve_st1##NAME##_host, sve_st1##NAME##_tlb); \ +static inline QEMU_ALWAYS_INLINE +void sve_stN_r_mte(CPUARMState *env, uint64_t *vg, target_ulong addr, + uint32_t desc, const uintptr_t ra, + const int esz, const int msz, const int N, + 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; + } + + sve_stN_r(env, vg, addr, desc, ra, esz, msz, N, mtedesc, host_fn, tlb_fn, + N == 1 ? sve_cont_ldst_mte_check1 : sve_cont_ldst_mte_checkN); } -#define DO_STN_2(N, NAME, ESZ, MSZ) \ -void HELPER(sve_st##N##NAME##_le_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_stN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, N, \ - sve_st1##NAME##_le_host, sve_st1##NAME##_le_tlb); \ -} \ -void HELPER(sve_st##N##NAME##_be_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_stN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, N, \ - sve_st1##NAME##_be_host, sve_st1##NAME##_be_tlb); \ +#define DO_STN_1(N, NAME, ESZ) \ +void HELPER(sve_st##N##NAME##_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_stN_r(env, vg, addr, desc, GETPC(), ESZ, MO_8, N, 0, \ + sve_st1##NAME##_host, sve_st1##NAME##_tlb, NULL); \ +} \ +void HELPER(sve_st##N##NAME##_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_stN_r_mte(env, vg, addr, desc, GETPC(), ESZ, MO_8, N, \ + sve_st1##NAME##_host, sve_st1##NAME##_tlb); \ +} + +#define DO_STN_2(N, NAME, ESZ, MSZ) \ +void HELPER(sve_st##N##NAME##_le_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_stN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, N, 0, \ + sve_st1##NAME##_le_host, sve_st1##NAME##_le_tlb, NULL); \ +} \ +void HELPER(sve_st##N##NAME##_be_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_stN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, N, 0, \ + sve_st1##NAME##_be_host, sve_st1##NAME##_be_tlb, NULL); \ +} \ +void HELPER(sve_st##N##NAME##_le_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_stN_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, N, \ + sve_st1##NAME##_le_host, sve_st1##NAME##_le_tlb); \ +} \ +void HELPER(sve_st##N##NAME##_be_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_stN_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, N, \ + sve_st1##NAME##_be_host, sve_st1##NAME##_be_tlb); \ } DO_STN_1(1, bb, MO_8) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_LD1R_zpri(DisasContext *s, arg_rpri_load *a) static void do_st_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, int msz, int esz, int nreg) { - static gen_helper_gvec_mem * const fn_single[2][4][4] = { - { { gen_helper_sve_st1bb_r, - gen_helper_sve_st1bh_r, - gen_helper_sve_st1bs_r, - gen_helper_sve_st1bd_r }, - { NULL, - gen_helper_sve_st1hh_le_r, - gen_helper_sve_st1hs_le_r, - gen_helper_sve_st1hd_le_r }, - { NULL, NULL, - gen_helper_sve_st1ss_le_r, - gen_helper_sve_st1sd_le_r }, - { NULL, NULL, NULL, - gen_helper_sve_st1dd_le_r } }, - { { gen_helper_sve_st1bb_r, - gen_helper_sve_st1bh_r, - gen_helper_sve_st1bs_r, - gen_helper_sve_st1bd_r }, - { NULL, - gen_helper_sve_st1hh_be_r, - gen_helper_sve_st1hs_be_r, - gen_helper_sve_st1hd_be_r }, - { NULL, NULL, - gen_helper_sve_st1ss_be_r, - gen_helper_sve_st1sd_be_r }, - { NULL, NULL, NULL, - gen_helper_sve_st1dd_be_r } }, + static gen_helper_gvec_mem * const fn_single[2][2][4][4] = { + { { { gen_helper_sve_st1bb_r, + gen_helper_sve_st1bh_r, + gen_helper_sve_st1bs_r, + gen_helper_sve_st1bd_r }, + { NULL, + gen_helper_sve_st1hh_le_r, + gen_helper_sve_st1hs_le_r, + gen_helper_sve_st1hd_le_r }, + { NULL, NULL, + gen_helper_sve_st1ss_le_r, + gen_helper_sve_st1sd_le_r }, + { NULL, NULL, NULL, + gen_helper_sve_st1dd_le_r } }, + { { gen_helper_sve_st1bb_r, + gen_helper_sve_st1bh_r, + gen_helper_sve_st1bs_r, + gen_helper_sve_st1bd_r }, + { NULL, + gen_helper_sve_st1hh_be_r, + gen_helper_sve_st1hs_be_r, + gen_helper_sve_st1hd_be_r }, + { NULL, NULL, + gen_helper_sve_st1ss_be_r, + gen_helper_sve_st1sd_be_r }, + { NULL, NULL, NULL, + gen_helper_sve_st1dd_be_r } } }, + + { { { gen_helper_sve_st1bb_r_mte, + gen_helper_sve_st1bh_r_mte, + gen_helper_sve_st1bs_r_mte, + gen_helper_sve_st1bd_r_mte }, + { NULL, + gen_helper_sve_st1hh_le_r_mte, + gen_helper_sve_st1hs_le_r_mte, + gen_helper_sve_st1hd_le_r_mte }, + { NULL, NULL, + gen_helper_sve_st1ss_le_r_mte, + gen_helper_sve_st1sd_le_r_mte }, + { NULL, NULL, NULL, + gen_helper_sve_st1dd_le_r_mte } }, + { { gen_helper_sve_st1bb_r_mte, + gen_helper_sve_st1bh_r_mte, + gen_helper_sve_st1bs_r_mte, + gen_helper_sve_st1bd_r_mte }, + { NULL, + gen_helper_sve_st1hh_be_r_mte, + gen_helper_sve_st1hs_be_r_mte, + gen_helper_sve_st1hd_be_r_mte }, + { NULL, NULL, + gen_helper_sve_st1ss_be_r_mte, + gen_helper_sve_st1sd_be_r_mte }, + { NULL, NULL, NULL, + gen_helper_sve_st1dd_be_r_mte } } }, }; - static gen_helper_gvec_mem * const fn_multiple[2][3][4] = { - { { gen_helper_sve_st2bb_r, - gen_helper_sve_st2hh_le_r, - gen_helper_sve_st2ss_le_r, - gen_helper_sve_st2dd_le_r }, - { gen_helper_sve_st3bb_r, - gen_helper_sve_st3hh_le_r, - gen_helper_sve_st3ss_le_r, - gen_helper_sve_st3dd_le_r }, - { gen_helper_sve_st4bb_r, - gen_helper_sve_st4hh_le_r, - gen_helper_sve_st4ss_le_r, - gen_helper_sve_st4dd_le_r } }, - { { gen_helper_sve_st2bb_r, - gen_helper_sve_st2hh_be_r, - gen_helper_sve_st2ss_be_r, - gen_helper_sve_st2dd_be_r }, - { gen_helper_sve_st3bb_r, - gen_helper_sve_st3hh_be_r, - gen_helper_sve_st3ss_be_r, - gen_helper_sve_st3dd_be_r }, - { gen_helper_sve_st4bb_r, - gen_helper_sve_st4hh_be_r, - gen_helper_sve_st4ss_be_r, - gen_helper_sve_st4dd_be_r } }, + static gen_helper_gvec_mem * const fn_multiple[2][2][3][4] = { + { { { gen_helper_sve_st2bb_r, + gen_helper_sve_st2hh_le_r, + gen_helper_sve_st2ss_le_r, + gen_helper_sve_st2dd_le_r }, + { gen_helper_sve_st3bb_r, + gen_helper_sve_st3hh_le_r, + gen_helper_sve_st3ss_le_r, + gen_helper_sve_st3dd_le_r }, + { gen_helper_sve_st4bb_r, + gen_helper_sve_st4hh_le_r, + gen_helper_sve_st4ss_le_r, + gen_helper_sve_st4dd_le_r } }, + { { gen_helper_sve_st2bb_r, + gen_helper_sve_st2hh_be_r, + gen_helper_sve_st2ss_be_r, + gen_helper_sve_st2dd_be_r }, + { gen_helper_sve_st3bb_r, + gen_helper_sve_st3hh_be_r, + gen_helper_sve_st3ss_be_r, + gen_helper_sve_st3dd_be_r }, + { gen_helper_sve_st4bb_r, + gen_helper_sve_st4hh_be_r, + gen_helper_sve_st4ss_be_r, + gen_helper_sve_st4dd_be_r } } }, + { { { gen_helper_sve_st2bb_r_mte, + gen_helper_sve_st2hh_le_r_mte, + gen_helper_sve_st2ss_le_r_mte, + gen_helper_sve_st2dd_le_r_mte }, + { gen_helper_sve_st3bb_r_mte, + gen_helper_sve_st3hh_le_r_mte, + gen_helper_sve_st3ss_le_r_mte, + gen_helper_sve_st3dd_le_r_mte }, + { gen_helper_sve_st4bb_r_mte, + gen_helper_sve_st4hh_le_r_mte, + gen_helper_sve_st4ss_le_r_mte, + gen_helper_sve_st4dd_le_r_mte } }, + { { gen_helper_sve_st2bb_r_mte, + gen_helper_sve_st2hh_be_r_mte, + gen_helper_sve_st2ss_be_r_mte, + gen_helper_sve_st2dd_be_r_mte }, + { gen_helper_sve_st3bb_r_mte, + gen_helper_sve_st3hh_be_r_mte, + gen_helper_sve_st3ss_be_r_mte, + gen_helper_sve_st3dd_be_r_mte }, + { gen_helper_sve_st4bb_r_mte, + gen_helper_sve_st4hh_be_r_mte, + gen_helper_sve_st4ss_be_r_mte, + gen_helper_sve_st4dd_be_r_mte } } }, }; gen_helper_gvec_mem *fn; int be = s->be_data == MO_BE; if (nreg == 0) { /* ST1 */ - fn = fn_single[be][msz][esz]; + fn = fn_single[s->mte_active[0]][be][msz][esz]; + nreg = 1; } else { /* ST2, ST3, ST4 -- msz == esz, enforced by encoding */ assert(msz == esz); - fn = fn_multiple[be][nreg - 1][msz]; + fn = fn_multiple[s->mte_active[0]][be][nreg - 1][msz]; } assert(fn != NULL); - do_mem_zpa(s, zt, pg, addr, msz_dtype(s, msz), 0, true, fn); + do_mem_zpa(s, zt, pg, addr, msz_dtype(s, msz), nreg, true, fn); } static bool trans_ST_zprr(DisasContext *s, arg_rprr_store *a) -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Because the elements are sequential, we can eliminate many tests all at once when the tag hits TCMA, or if the page(s) are not Tagged. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-36-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sve.h | 98 ++++++++++++++++ target/arm/sve_helper.c | 99 ++++++++++++++-- target/arm/translate-sve.c | 232 +++++++++++++++++++++++++------------ 3 files changed, 343 insertions(+), 86 deletions(-) diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sve_ldff1sds_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldff1dd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldff1dd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1bhu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1bsu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1bdu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1bhs_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1bss_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1bds_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldff1hh_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hsu_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hdu_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hss_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hds_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldff1hh_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hsu_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hdu_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hss_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hds_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldff1ss_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1sdu_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1sds_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldff1ss_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1sdu_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1sds_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldff1dd_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1dd_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + DEF_HELPER_FLAGS_4(sve_ldnf1bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldnf1bhu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldnf1bsu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sve_ldnf1sds_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldnf1dd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldnf1dd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1bhu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1bsu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1bdu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1bhs_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1bss_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1bds_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldnf1hh_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hsu_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hdu_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hss_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hds_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldnf1hh_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hsu_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hdu_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hss_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hds_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldnf1ss_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1sdu_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1sds_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldnf1ss_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1sdu_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1sds_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldnf1dd_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1dd_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + DEF_HELPER_FLAGS_4(sve_st1bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st2bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st3bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ static void record_fault(CPUARMState *env, uintptr_t i, uintptr_t oprsz) */ static inline QEMU_ALWAYS_INLINE void sve_ldnfff1_r(CPUARMState *env, void *vg, const target_ulong addr, - uint32_t desc, const uintptr_t retaddr, + uint32_t desc, const uintptr_t retaddr, uint32_t mtedesc, const int esz, const int msz, const SVEContFault fault, sve_ldst1_host_fn *host_fn, sve_ldst1_tlb_fn *tlb_fn) @@ -XXX,XX +XXX,XX @@ void sve_ldnfff1_r(CPUARMState *env, void *vg, const target_ulong addr, mem_off = info.mem_off_first[0]; flags = info.page[0].flags; + /* + * Disable MTE checking if the Tagged bit is not set. Since TBI must + * be set within MTEDESC for MTE, !mtedesc => !mte_active. + */ + if (arm_tlb_mte_tagged(&info.page[0].attrs)) { + mtedesc = 0; + } + if (fault == FAULT_FIRST) { + /* Trapping mte check for the first-fault element. */ + if (mtedesc) { + mte_check1(env, mtedesc, addr + mem_off, retaddr); + } + /* * Special handling of the first active element, * if it crosses a page boundary or is MMIO. */ bool is_split = mem_off == info.mem_off_split; - /* TODO: MTE check. */ if (unlikely(flags != 0) || unlikely(is_split)) { /* * Use the slow path for cross-page handling. @@ -XXX,XX +XXX,XX @@ void sve_ldnfff1_r(CPUARMState *env, void *vg, const target_ulong addr, /* Watchpoint hit, see below. */ goto do_fault; } - /* TODO: MTE check. */ + if (mtedesc && !mte_probe1(env, mtedesc, addr + mem_off)) { + goto do_fault; + } /* * Use the slow path for cross-page handling. * This is RAM, without a watchpoint, and will not trap. @@ -XXX,XX +XXX,XX @@ void sve_ldnfff1_r(CPUARMState *env, void *vg, const target_ulong addr, & BP_MEM_READ)) { goto do_fault; } - /* TODO: MTE check. */ + if (mtedesc && !mte_probe1(env, mtedesc, addr + mem_off)) { + goto do_fault; + } host_fn(vd, reg_off, host + mem_off); } reg_off += 1 << esz; @@ -XXX,XX +XXX,XX @@ void sve_ldnfff1_r(CPUARMState *env, void *vg, const target_ulong addr, record_fault(env, reg_off, reg_max); } -#define DO_LDFF1_LDNF1_1(PART, ESZ) \ +static inline QEMU_ALWAYS_INLINE +void sve_ldnfff1_r_mte(CPUARMState *env, void *vg, target_ulong addr, + uint32_t desc, const uintptr_t retaddr, + const int esz, const int msz, const SVEContFault fault, + 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; + } + + sve_ldnfff1_r(env, vg, addr, desc, retaddr, mtedesc, + esz, msz, fault, host_fn, tlb_fn); +} + +#define DO_LDFF1_LDNF1_1(PART, ESZ) \ void HELPER(sve_ldff1##PART##_r)(CPUARMState *env, void *vg, \ target_ulong addr, uint32_t desc) \ { \ - sve_ldnfff1_r(env, vg, addr, desc, GETPC(), ESZ, MO_8, FAULT_FIRST, \ + sve_ldnfff1_r(env, vg, addr, desc, GETPC(), 0, ESZ, MO_8, FAULT_FIRST, \ sve_ld1##PART##_host, sve_ld1##PART##_tlb); \ } \ void HELPER(sve_ldnf1##PART##_r)(CPUARMState *env, void *vg, \ target_ulong addr, uint32_t desc) \ { \ - sve_ldnfff1_r(env, vg, addr, desc, GETPC(), ESZ, MO_8, FAULT_NO, \ + sve_ldnfff1_r(env, vg, addr, desc, GETPC(), 0, ESZ, MO_8, FAULT_NO, \ + sve_ld1##PART##_host, sve_ld1##PART##_tlb); \ +} \ +void HELPER(sve_ldff1##PART##_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldnfff1_r_mte(env, vg, addr, desc, GETPC(), ESZ, MO_8, FAULT_FIRST, \ + sve_ld1##PART##_host, sve_ld1##PART##_tlb); \ +} \ +void HELPER(sve_ldnf1##PART##_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldnfff1_r_mte(env, vg, addr, desc, GETPC(), ESZ, MO_8, FAULT_NO, \ sve_ld1##PART##_host, sve_ld1##PART##_tlb); \ } -#define DO_LDFF1_LDNF1_2(PART, ESZ, MSZ) \ +#define DO_LDFF1_LDNF1_2(PART, ESZ, MSZ) \ void HELPER(sve_ldff1##PART##_le_r)(CPUARMState *env, void *vg, \ target_ulong addr, uint32_t desc) \ { \ - sve_ldnfff1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_FIRST, \ + sve_ldnfff1_r(env, vg, addr, desc, GETPC(), 0, ESZ, MSZ, FAULT_FIRST, \ sve_ld1##PART##_le_host, sve_ld1##PART##_le_tlb); \ } \ void HELPER(sve_ldnf1##PART##_le_r)(CPUARMState *env, void *vg, \ target_ulong addr, uint32_t desc) \ { \ - sve_ldnfff1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_NO, \ + sve_ldnfff1_r(env, vg, addr, desc, GETPC(), 0, ESZ, MSZ, FAULT_NO, \ sve_ld1##PART##_le_host, sve_ld1##PART##_le_tlb); \ } \ void HELPER(sve_ldff1##PART##_be_r)(CPUARMState *env, void *vg, \ target_ulong addr, uint32_t desc) \ { \ - sve_ldnfff1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_FIRST, \ + sve_ldnfff1_r(env, vg, addr, desc, GETPC(), 0, ESZ, MSZ, FAULT_FIRST, \ sve_ld1##PART##_be_host, sve_ld1##PART##_be_tlb); \ } \ void HELPER(sve_ldnf1##PART##_be_r)(CPUARMState *env, void *vg, \ target_ulong addr, uint32_t desc) \ { \ - sve_ldnfff1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_NO, \ + sve_ldnfff1_r(env, vg, addr, desc, GETPC(), 0, ESZ, MSZ, FAULT_NO, \ sve_ld1##PART##_be_host, sve_ld1##PART##_be_tlb); \ +} \ +void HELPER(sve_ldff1##PART##_le_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldnfff1_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_FIRST, \ + sve_ld1##PART##_le_host, sve_ld1##PART##_le_tlb); \ +} \ +void HELPER(sve_ldnf1##PART##_le_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldnfff1_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_NO, \ + sve_ld1##PART##_le_host, sve_ld1##PART##_le_tlb); \ +} \ +void HELPER(sve_ldff1##PART##_be_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldnfff1_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_FIRST, \ + sve_ld1##PART##_be_host, sve_ld1##PART##_be_tlb); \ +} \ +void HELPER(sve_ldnf1##PART##_be_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldnfff1_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_NO, \ + sve_ld1##PART##_be_host, sve_ld1##PART##_be_tlb); \ } DO_LDFF1_LDNF1_1(bb, MO_8) 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_LD_zpri(DisasContext *s, arg_rpri_load *a) static bool trans_LDFF1_zprr(DisasContext *s, arg_rprr_load *a) { - static gen_helper_gvec_mem * const fns[2][16] = { - /* Little-endian */ - { gen_helper_sve_ldff1bb_r, - gen_helper_sve_ldff1bhu_r, - gen_helper_sve_ldff1bsu_r, - gen_helper_sve_ldff1bdu_r, + static gen_helper_gvec_mem * const fns[2][2][16] = { + { /* mte inactive, little-endian */ + { gen_helper_sve_ldff1bb_r, + gen_helper_sve_ldff1bhu_r, + gen_helper_sve_ldff1bsu_r, + gen_helper_sve_ldff1bdu_r, - gen_helper_sve_ldff1sds_le_r, - gen_helper_sve_ldff1hh_le_r, - gen_helper_sve_ldff1hsu_le_r, - gen_helper_sve_ldff1hdu_le_r, + gen_helper_sve_ldff1sds_le_r, + gen_helper_sve_ldff1hh_le_r, + gen_helper_sve_ldff1hsu_le_r, + gen_helper_sve_ldff1hdu_le_r, - gen_helper_sve_ldff1hds_le_r, - gen_helper_sve_ldff1hss_le_r, - gen_helper_sve_ldff1ss_le_r, - gen_helper_sve_ldff1sdu_le_r, + gen_helper_sve_ldff1hds_le_r, + gen_helper_sve_ldff1hss_le_r, + gen_helper_sve_ldff1ss_le_r, + gen_helper_sve_ldff1sdu_le_r, - gen_helper_sve_ldff1bds_r, - gen_helper_sve_ldff1bss_r, - gen_helper_sve_ldff1bhs_r, - gen_helper_sve_ldff1dd_le_r }, + gen_helper_sve_ldff1bds_r, + gen_helper_sve_ldff1bss_r, + gen_helper_sve_ldff1bhs_r, + gen_helper_sve_ldff1dd_le_r }, - /* Big-endian */ - { gen_helper_sve_ldff1bb_r, - gen_helper_sve_ldff1bhu_r, - gen_helper_sve_ldff1bsu_r, - gen_helper_sve_ldff1bdu_r, + /* mte inactive, big-endian */ + { gen_helper_sve_ldff1bb_r, + gen_helper_sve_ldff1bhu_r, + gen_helper_sve_ldff1bsu_r, + gen_helper_sve_ldff1bdu_r, - gen_helper_sve_ldff1sds_be_r, - gen_helper_sve_ldff1hh_be_r, - gen_helper_sve_ldff1hsu_be_r, - gen_helper_sve_ldff1hdu_be_r, + gen_helper_sve_ldff1sds_be_r, + gen_helper_sve_ldff1hh_be_r, + gen_helper_sve_ldff1hsu_be_r, + gen_helper_sve_ldff1hdu_be_r, - gen_helper_sve_ldff1hds_be_r, - gen_helper_sve_ldff1hss_be_r, - gen_helper_sve_ldff1ss_be_r, - gen_helper_sve_ldff1sdu_be_r, + gen_helper_sve_ldff1hds_be_r, + gen_helper_sve_ldff1hss_be_r, + gen_helper_sve_ldff1ss_be_r, + gen_helper_sve_ldff1sdu_be_r, - gen_helper_sve_ldff1bds_r, - gen_helper_sve_ldff1bss_r, - gen_helper_sve_ldff1bhs_r, - gen_helper_sve_ldff1dd_be_r }, + gen_helper_sve_ldff1bds_r, + gen_helper_sve_ldff1bss_r, + gen_helper_sve_ldff1bhs_r, + gen_helper_sve_ldff1dd_be_r } }, + + { /* mte active, little-endian */ + { gen_helper_sve_ldff1bb_r_mte, + gen_helper_sve_ldff1bhu_r_mte, + gen_helper_sve_ldff1bsu_r_mte, + gen_helper_sve_ldff1bdu_r_mte, + + gen_helper_sve_ldff1sds_le_r_mte, + gen_helper_sve_ldff1hh_le_r_mte, + gen_helper_sve_ldff1hsu_le_r_mte, + gen_helper_sve_ldff1hdu_le_r_mte, + + gen_helper_sve_ldff1hds_le_r_mte, + gen_helper_sve_ldff1hss_le_r_mte, + gen_helper_sve_ldff1ss_le_r_mte, + gen_helper_sve_ldff1sdu_le_r_mte, + + gen_helper_sve_ldff1bds_r_mte, + gen_helper_sve_ldff1bss_r_mte, + gen_helper_sve_ldff1bhs_r_mte, + gen_helper_sve_ldff1dd_le_r_mte }, + + /* mte active, big-endian */ + { gen_helper_sve_ldff1bb_r_mte, + gen_helper_sve_ldff1bhu_r_mte, + gen_helper_sve_ldff1bsu_r_mte, + gen_helper_sve_ldff1bdu_r_mte, + + gen_helper_sve_ldff1sds_be_r_mte, + gen_helper_sve_ldff1hh_be_r_mte, + gen_helper_sve_ldff1hsu_be_r_mte, + gen_helper_sve_ldff1hdu_be_r_mte, + + gen_helper_sve_ldff1hds_be_r_mte, + gen_helper_sve_ldff1hss_be_r_mte, + gen_helper_sve_ldff1ss_be_r_mte, + gen_helper_sve_ldff1sdu_be_r_mte, + + gen_helper_sve_ldff1bds_r_mte, + gen_helper_sve_ldff1bss_r_mte, + gen_helper_sve_ldff1bhs_r_mte, + gen_helper_sve_ldff1dd_be_r_mte } }, }; if (sve_access_check(s)) { TCGv_i64 addr = new_tmp_a64(s); tcg_gen_shli_i64(addr, cpu_reg(s, a->rm), dtype_msz(a->dtype)); tcg_gen_add_i64(addr, addr, cpu_reg_sp(s, a->rn)); - do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, 0, false, - fns[s->be_data == MO_BE][a->dtype]); + do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, 1, false, + fns[s->mte_active[0]][s->be_data == MO_BE][a->dtype]); } return true; } static bool trans_LDNF1_zpri(DisasContext *s, arg_rpri_load *a) { - static gen_helper_gvec_mem * const fns[2][16] = { - /* Little-endian */ - { gen_helper_sve_ldnf1bb_r, - gen_helper_sve_ldnf1bhu_r, - gen_helper_sve_ldnf1bsu_r, - gen_helper_sve_ldnf1bdu_r, + static gen_helper_gvec_mem * const fns[2][2][16] = { + { /* mte inactive, little-endian */ + { gen_helper_sve_ldnf1bb_r, + gen_helper_sve_ldnf1bhu_r, + gen_helper_sve_ldnf1bsu_r, + gen_helper_sve_ldnf1bdu_r, - gen_helper_sve_ldnf1sds_le_r, - gen_helper_sve_ldnf1hh_le_r, - gen_helper_sve_ldnf1hsu_le_r, - gen_helper_sve_ldnf1hdu_le_r, + gen_helper_sve_ldnf1sds_le_r, + gen_helper_sve_ldnf1hh_le_r, + gen_helper_sve_ldnf1hsu_le_r, + gen_helper_sve_ldnf1hdu_le_r, - gen_helper_sve_ldnf1hds_le_r, - gen_helper_sve_ldnf1hss_le_r, - gen_helper_sve_ldnf1ss_le_r, - gen_helper_sve_ldnf1sdu_le_r, + gen_helper_sve_ldnf1hds_le_r, + gen_helper_sve_ldnf1hss_le_r, + gen_helper_sve_ldnf1ss_le_r, + gen_helper_sve_ldnf1sdu_le_r, - gen_helper_sve_ldnf1bds_r, - gen_helper_sve_ldnf1bss_r, - gen_helper_sve_ldnf1bhs_r, - gen_helper_sve_ldnf1dd_le_r }, + gen_helper_sve_ldnf1bds_r, + gen_helper_sve_ldnf1bss_r, + gen_helper_sve_ldnf1bhs_r, + gen_helper_sve_ldnf1dd_le_r }, - /* Big-endian */ - { gen_helper_sve_ldnf1bb_r, - gen_helper_sve_ldnf1bhu_r, - gen_helper_sve_ldnf1bsu_r, - gen_helper_sve_ldnf1bdu_r, + /* mte inactive, big-endian */ + { gen_helper_sve_ldnf1bb_r, + gen_helper_sve_ldnf1bhu_r, + gen_helper_sve_ldnf1bsu_r, + gen_helper_sve_ldnf1bdu_r, - gen_helper_sve_ldnf1sds_be_r, - gen_helper_sve_ldnf1hh_be_r, - gen_helper_sve_ldnf1hsu_be_r, - gen_helper_sve_ldnf1hdu_be_r, + gen_helper_sve_ldnf1sds_be_r, + gen_helper_sve_ldnf1hh_be_r, + gen_helper_sve_ldnf1hsu_be_r, + gen_helper_sve_ldnf1hdu_be_r, - gen_helper_sve_ldnf1hds_be_r, - gen_helper_sve_ldnf1hss_be_r, - gen_helper_sve_ldnf1ss_be_r, - gen_helper_sve_ldnf1sdu_be_r, + gen_helper_sve_ldnf1hds_be_r, + gen_helper_sve_ldnf1hss_be_r, + gen_helper_sve_ldnf1ss_be_r, + gen_helper_sve_ldnf1sdu_be_r, - gen_helper_sve_ldnf1bds_r, - gen_helper_sve_ldnf1bss_r, - gen_helper_sve_ldnf1bhs_r, - gen_helper_sve_ldnf1dd_be_r }, + gen_helper_sve_ldnf1bds_r, + gen_helper_sve_ldnf1bss_r, + gen_helper_sve_ldnf1bhs_r, + gen_helper_sve_ldnf1dd_be_r } }, + + { /* mte inactive, little-endian */ + { gen_helper_sve_ldnf1bb_r_mte, + gen_helper_sve_ldnf1bhu_r_mte, + gen_helper_sve_ldnf1bsu_r_mte, + gen_helper_sve_ldnf1bdu_r_mte, + + gen_helper_sve_ldnf1sds_le_r_mte, + gen_helper_sve_ldnf1hh_le_r_mte, + gen_helper_sve_ldnf1hsu_le_r_mte, + gen_helper_sve_ldnf1hdu_le_r_mte, + + gen_helper_sve_ldnf1hds_le_r_mte, + gen_helper_sve_ldnf1hss_le_r_mte, + gen_helper_sve_ldnf1ss_le_r_mte, + gen_helper_sve_ldnf1sdu_le_r_mte, + + gen_helper_sve_ldnf1bds_r_mte, + gen_helper_sve_ldnf1bss_r_mte, + gen_helper_sve_ldnf1bhs_r_mte, + gen_helper_sve_ldnf1dd_le_r_mte }, + + /* mte inactive, big-endian */ + { gen_helper_sve_ldnf1bb_r_mte, + gen_helper_sve_ldnf1bhu_r_mte, + gen_helper_sve_ldnf1bsu_r_mte, + gen_helper_sve_ldnf1bdu_r_mte, + + gen_helper_sve_ldnf1sds_be_r_mte, + gen_helper_sve_ldnf1hh_be_r_mte, + gen_helper_sve_ldnf1hsu_be_r_mte, + gen_helper_sve_ldnf1hdu_be_r_mte, + + gen_helper_sve_ldnf1hds_be_r_mte, + gen_helper_sve_ldnf1hss_be_r_mte, + gen_helper_sve_ldnf1ss_be_r_mte, + gen_helper_sve_ldnf1sdu_be_r_mte, + + gen_helper_sve_ldnf1bds_r_mte, + gen_helper_sve_ldnf1bss_r_mte, + gen_helper_sve_ldnf1bhs_r_mte, + gen_helper_sve_ldnf1dd_be_r_mte } }, }; if (sve_access_check(s)) { @@ -XXX,XX +XXX,XX @@ static bool trans_LDNF1_zpri(DisasContext *s, arg_rpri_load *a) TCGv_i64 addr = new_tmp_a64(s); tcg_gen_addi_i64(addr, cpu_reg_sp(s, a->rn), off); - do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, 0, false, - fns[s->be_data == MO_BE][a->dtype]); + do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, 1, false, + fns[s->mte_active[0]][s->be_data == MO_BE][a->dtype]); } return true; } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> We still need to handle tbi for user-only when mte is inactive. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-37-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.h | 1 + target/arm/translate-a64.c | 2 +- target/arm/translate-sve.c | 6 ++++-- 3 files changed, 6 insertions(+), 3 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 @@ TCGv_ptr get_fpstatus_ptr(bool); bool logic_imm_decode_wmask(uint64_t *result, unsigned int immn, unsigned int imms, unsigned int immr); bool sve_access_check(DisasContext *s); +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); TCGv_i64 gen_mte_checkN(DisasContext *s, TCGv_i64 addr, bool is_write, diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void gen_a64_set_pc(DisasContext *s, TCGv_i64 src) * of the write-back address. */ -static TCGv_i64 clean_data_tbi(DisasContext *s, TCGv_i64 addr) +TCGv_i64 clean_data_tbi(DisasContext *s, TCGv_i64 addr) { TCGv_i64 clean = new_tmp_a64(s); #ifdef CONFIG_USER_ONLY diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static void do_mem_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, * For e.g. LD4, there are not enough arguments to pass all 4 * registers as pointers, so encode the regno into the data field. * For consistency, do this even for LD1. - * TODO: mte_n check here while callers are updated. */ - if (mte_n && s->mte_active[0]) { + if (s->mte_active[0]) { int msz = dtype_msz(dtype); desc = FIELD_DP32(desc, MTEDESC, MIDX, get_mem_index(s)); @@ -XXX,XX +XXX,XX @@ static void do_mem_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, desc = FIELD_DP32(desc, MTEDESC, ESIZE, 1 << msz); desc = FIELD_DP32(desc, MTEDESC, TSIZE, mte_n << msz); desc <<= SVE_MTEDESC_SHIFT; + } else { + addr = clean_data_tbi(s, addr); } + desc = simd_desc(vsz, vsz, zt | desc); t_desc = tcg_const_i32(desc); t_pg = tcg_temp_new_ptr(); -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Because the elements are non-sequential, we cannot eliminate many tests straight away like we can for sequential operations. But we often have the PTE details handy, so we can test for Tagged. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-38-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sve.h | 285 ++++++++++++++++ target/arm/sve_helper.c | 185 +++++++++-- target/arm/translate-sve.c | 650 +++++++++++++++++++++++++------------ 3 files changed, 872 insertions(+), 248 deletions(-) diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_6(sve_ldsds_le_zd, TCG_CALL_NO_WG, DEF_HELPER_FLAGS_6(sve_ldsds_be_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldbsu_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhsu_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhsu_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldss_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldss_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldbss_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhss_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhss_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldbsu_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhsu_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhsu_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldss_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldss_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldbss_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhss_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhss_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldbdu_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhdu_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhdu_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldbds_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhds_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhds_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldbdu_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhdu_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhdu_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldbds_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhds_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhds_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldbdu_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhdu_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhdu_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldbds_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhds_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhds_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + DEF_HELPER_FLAGS_6(sve_ldffbsu_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldffhsu_le_zsu, TCG_CALL_NO_WG, @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_6(sve_ldffsds_le_zd, TCG_CALL_NO_WG, DEF_HELPER_FLAGS_6(sve_ldffsds_be_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffbsu_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhsu_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhsu_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffss_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffss_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffbss_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhss_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhss_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldffbsu_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhsu_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhsu_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffss_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffss_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffbss_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhss_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhss_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldffbdu_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhdu_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhdu_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffbds_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhds_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhds_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldffbdu_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhdu_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhdu_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffbds_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhds_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhds_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldffbdu_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhdu_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhdu_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffbds_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhds_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhds_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + DEF_HELPER_FLAGS_6(sve_stbs_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_sths_le_zsu, TCG_CALL_NO_WG, @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_6(sve_stdd_le_zd, TCG_CALL_NO_WG, DEF_HELPER_FLAGS_6(sve_stdd_be_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stbs_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sths_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sths_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stss_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stss_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_stbs_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sths_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sths_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stss_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stss_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_stbd_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sthd_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sthd_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_stbd_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sthd_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sthd_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_stbd_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sthd_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sthd_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + DEF_HELPER_FLAGS_4(sve2_pmull_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ static target_ulong off_zd_d(void *reg, intptr_t reg_ofs) static inline QEMU_ALWAYS_INLINE void sve_ld1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, target_ulong base, uint32_t desc, uintptr_t retaddr, - int esize, int msize, zreg_off_fn *off_fn, + uint32_t mtedesc, int esize, int msize, + zreg_off_fn *off_fn, sve_ldst1_host_fn *host_fn, sve_ldst1_tlb_fn *tlb_fn) { @@ -XXX,XX +XXX,XX @@ void sve_ld1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, cpu_check_watchpoint(env_cpu(env), addr, msize, info.attrs, BP_MEM_READ, retaddr); } - /* TODO: MTE check */ + if (mtedesc && arm_tlb_mte_tagged(&info.attrs)) { + mte_check1(env, mtedesc, addr, retaddr); + } host_fn(&scratch, reg_off, info.host); } else { /* Element crosses the page boundary. */ @@ -XXX,XX +XXX,XX @@ void sve_ld1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, msize, info.attrs, BP_MEM_READ, retaddr); } - /* TODO: MTE check */ + if (mtedesc && arm_tlb_mte_tagged(&info.attrs)) { + mte_check1(env, mtedesc, addr, retaddr); + } tlb_fn(env, &scratch, reg_off, addr, retaddr); } } @@ -XXX,XX +XXX,XX @@ void sve_ld1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, memcpy(vd, &scratch, reg_max); } +static inline QEMU_ALWAYS_INLINE +void sve_ld1_z_mte(CPUARMState *env, void *vd, uint64_t *vg, void *vm, + target_ulong base, uint32_t desc, uintptr_t retaddr, + int esize, int msize, zreg_off_fn *off_fn, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn) +{ + uint32_t mtedesc = desc >> (SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + /* Remove mtedesc from the normal sve descriptor. */ + desc = extract32(desc, 0, SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + + /* + * ??? TODO: For the 32-bit offset extractions, base + ofs cannot + * offset base entirely over the address space hole to change the + * pointer tag, or change the bit55 selector. So we could here + * examine TBI + TCMA like we do for sve_ldN_r_mte(). + */ + sve_ld1_z(env, vd, vg, vm, base, desc, retaddr, mtedesc, + esize, msize, off_fn, host_fn, tlb_fn); +} + #define DO_LD1_ZPZ_S(MEM, OFS, MSZ) \ void HELPER(sve_ld##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ void *vm, target_ulong base, uint32_t desc) \ { \ - sve_ld1_z(env, vd, vg, vm, base, desc, GETPC(), 4, 1 << MSZ, \ + sve_ld1_z(env, vd, vg, vm, base, desc, GETPC(), 0, 4, 1 << MSZ, \ off_##OFS##_s, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ +} \ +void HELPER(sve_ld##MEM##_##OFS##_mte)(CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_ld1_z_mte(env, vd, vg, vm, base, desc, GETPC(), 4, 1 << MSZ, \ + off_##OFS##_s, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ } #define DO_LD1_ZPZ_D(MEM, OFS, MSZ) \ void HELPER(sve_ld##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ void *vm, target_ulong base, uint32_t desc) \ { \ - sve_ld1_z(env, vd, vg, vm, base, desc, GETPC(), 8, 1 << MSZ, \ + sve_ld1_z(env, vd, vg, vm, base, desc, GETPC(), 0, 8, 1 << MSZ, \ off_##OFS##_d, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ +} \ +void HELPER(sve_ld##MEM##_##OFS##_mte)(CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_ld1_z_mte(env, vd, vg, vm, base, desc, GETPC(), 8, 1 << MSZ, \ + off_##OFS##_d, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ } DO_LD1_ZPZ_S(bsu, zsu, MO_8) @@ -XXX,XX +XXX,XX @@ DO_LD1_ZPZ_D(dd_be, zd, MO_64) static inline QEMU_ALWAYS_INLINE void sve_ldff1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, target_ulong base, uint32_t desc, uintptr_t retaddr, - const int esz, const int msz, zreg_off_fn *off_fn, + uint32_t mtedesc, const int esz, const int msz, + zreg_off_fn *off_fn, sve_ldst1_host_fn *host_fn, sve_ldst1_tlb_fn *tlb_fn) { @@ -XXX,XX +XXX,XX @@ void sve_ldff1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, * Probe the first element, allowing faults. */ addr = base + (off_fn(vm, reg_off) << scale); + if (mtedesc) { + mte_check1(env, mtedesc, addr, retaddr); + } tlb_fn(env, vd, reg_off, addr, retaddr); /* After any fault, zero the other elements. */ @@ -XXX,XX +XXX,XX @@ void sve_ldff1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, (env_cpu(env), addr, msize) & BP_MEM_READ)) { goto fault; } - /* TODO: MTE check. */ + if (mtedesc && + arm_tlb_mte_tagged(&info.attrs) && + !mte_probe1(env, mtedesc, addr)) { + goto fault; + } host_fn(vd, reg_off, info.host); } @@ -XXX,XX +XXX,XX @@ void sve_ldff1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, record_fault(env, reg_off, reg_max); } -#define DO_LDFF1_ZPZ_S(MEM, OFS, MSZ) \ -void HELPER(sve_ldff##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ - void *vm, target_ulong base, uint32_t desc) \ -{ \ - sve_ldff1_z(env, vd, vg, vm, base, desc, GETPC(), MO_32, MSZ, \ - off_##OFS##_s, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ +static inline QEMU_ALWAYS_INLINE +void sve_ldff1_z_mte(CPUARMState *env, void *vd, uint64_t *vg, void *vm, + target_ulong base, uint32_t desc, uintptr_t retaddr, + const int esz, const int msz, + zreg_off_fn *off_fn, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn) +{ + uint32_t mtedesc = desc >> (SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + /* Remove mtedesc from the normal sve descriptor. */ + desc = extract32(desc, 0, SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + + /* + * ??? TODO: For the 32-bit offset extractions, base + ofs cannot + * offset base entirely over the address space hole to change the + * pointer tag, or change the bit55 selector. So we could here + * examine TBI + TCMA like we do for sve_ldN_r_mte(). + */ + sve_ldff1_z(env, vd, vg, vm, base, desc, retaddr, mtedesc, + esz, msz, off_fn, host_fn, tlb_fn); } -#define DO_LDFF1_ZPZ_D(MEM, OFS, MSZ) \ -void HELPER(sve_ldff##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ - void *vm, target_ulong base, uint32_t desc) \ -{ \ - sve_ldff1_z(env, vd, vg, vm, base, desc, GETPC(), MO_64, MSZ, \ - off_##OFS##_d, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ +#define DO_LDFF1_ZPZ_S(MEM, OFS, MSZ) \ +void HELPER(sve_ldff##MEM##_##OFS) \ + (CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_ldff1_z(env, vd, vg, vm, base, desc, GETPC(), 0, MO_32, MSZ, \ + off_##OFS##_s, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ +} \ +void HELPER(sve_ldff##MEM##_##OFS##_mte) \ + (CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_ldff1_z_mte(env, vd, vg, vm, base, desc, GETPC(), MO_32, MSZ, \ + off_##OFS##_s, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ +} + +#define DO_LDFF1_ZPZ_D(MEM, OFS, MSZ) \ +void HELPER(sve_ldff##MEM##_##OFS) \ + (CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_ldff1_z(env, vd, vg, vm, base, desc, GETPC(), 0, MO_64, MSZ, \ + off_##OFS##_d, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ +} \ +void HELPER(sve_ldff##MEM##_##OFS##_mte) \ + (CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_ldff1_z_mte(env, vd, vg, vm, base, desc, GETPC(), MO_64, MSZ, \ + off_##OFS##_d, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ } DO_LDFF1_ZPZ_S(bsu, zsu, MO_8) @@ -XXX,XX +XXX,XX @@ DO_LDFF1_ZPZ_D(dd_be, zd, MO_64) static inline QEMU_ALWAYS_INLINE void sve_st1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, target_ulong base, uint32_t desc, uintptr_t retaddr, - int esize, int msize, zreg_off_fn *off_fn, + uint32_t mtedesc, int esize, int msize, + zreg_off_fn *off_fn, sve_ldst1_host_fn *host_fn, sve_ldst1_tlb_fn *tlb_fn) { @@ -XXX,XX +XXX,XX @@ void sve_st1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, cpu_check_watchpoint(env_cpu(env), addr, msize, info.attrs, BP_MEM_WRITE, retaddr); } - /* TODO: MTE check. */ + + if (mtedesc && arm_tlb_mte_tagged(&info.attrs)) { + mte_check1(env, mtedesc, addr, retaddr); + } } i += 1; reg_off += esize; @@ -XXX,XX +XXX,XX @@ void sve_st1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, } while (reg_off < reg_max); } -#define DO_ST1_ZPZ_S(MEM, OFS, MSZ) \ -void HELPER(sve_st##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ - void *vm, target_ulong base, uint32_t desc) \ -{ \ - sve_st1_z(env, vd, vg, vm, base, desc, GETPC(), 4, 1 << MSZ, \ - off_##OFS##_s, sve_st1##MEM##_host, sve_st1##MEM##_tlb); \ +static inline QEMU_ALWAYS_INLINE +void sve_st1_z_mte(CPUARMState *env, void *vd, uint64_t *vg, void *vm, + target_ulong base, uint32_t desc, uintptr_t retaddr, + int esize, int msize, zreg_off_fn *off_fn, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn) +{ + uint32_t mtedesc = desc >> (SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + /* Remove mtedesc from the normal sve descriptor. */ + desc = extract32(desc, 0, SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + + /* + * ??? TODO: For the 32-bit offset extractions, base + ofs cannot + * offset base entirely over the address space hole to change the + * pointer tag, or change the bit55 selector. So we could here + * examine TBI + TCMA like we do for sve_ldN_r_mte(). + */ + sve_st1_z(env, vd, vg, vm, base, desc, retaddr, mtedesc, + esize, msize, off_fn, host_fn, tlb_fn); } -#define DO_ST1_ZPZ_D(MEM, OFS, MSZ) \ -void HELPER(sve_st##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ +#define DO_ST1_ZPZ_S(MEM, OFS, MSZ) \ +void HELPER(sve_st##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ void *vm, target_ulong base, uint32_t desc) \ -{ \ - sve_st1_z(env, vd, vg, vm, base, desc, GETPC(), 8, 1 << MSZ, \ - off_##OFS##_d, sve_st1##MEM##_host, sve_st1##MEM##_tlb); \ +{ \ + sve_st1_z(env, vd, vg, vm, base, desc, GETPC(), 0, 4, 1 << MSZ, \ + off_##OFS##_s, sve_st1##MEM##_host, sve_st1##MEM##_tlb); \ +} \ +void HELPER(sve_st##MEM##_##OFS##_mte)(CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_st1_z_mte(env, vd, vg, vm, base, desc, GETPC(), 4, 1 << MSZ, \ + off_##OFS##_s, sve_st1##MEM##_host, sve_st1##MEM##_tlb); \ +} + +#define DO_ST1_ZPZ_D(MEM, OFS, MSZ) \ +void HELPER(sve_st##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_st1_z(env, vd, vg, vm, base, desc, GETPC(), 0, 8, 1 << MSZ, \ + off_##OFS##_d, sve_st1##MEM##_host, sve_st1##MEM##_tlb); \ +} \ +void HELPER(sve_st##MEM##_##OFS##_mte)(CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_st1_z_mte(env, vd, vg, vm, base, desc, GETPC(), 8, 1 << MSZ, \ + off_##OFS##_d, sve_st1##MEM##_host, sve_st1##MEM##_tlb); \ } DO_ST1_ZPZ_S(bs, zsu, MO_8) 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_ST_zpri(DisasContext *s, arg_rpri_store *a) */ static void do_mem_zpz(DisasContext *s, int zt, int pg, int zm, - int scale, TCGv_i64 scalar, int msz, + int scale, TCGv_i64 scalar, int msz, bool is_write, gen_helper_gvec_mem_scatter *fn) { unsigned vsz = vec_full_reg_size(s); @@ -XXX,XX +XXX,XX @@ static void do_mem_zpz(DisasContext *s, int zt, int pg, int zm, TCGv_ptr t_pg = tcg_temp_new_ptr(); TCGv_ptr t_zt = tcg_temp_new_ptr(); TCGv_i32 t_desc; - int desc; + int desc = 0; + if (s->mte_active[0]) { + 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, is_write); + desc = FIELD_DP32(desc, MTEDESC, ESIZE, 1 << msz); + desc <<= SVE_MTEDESC_SHIFT; + } desc = simd_desc(vsz, vsz, scale); t_desc = tcg_const_i32(desc); @@ -XXX,XX +XXX,XX @@ static void do_mem_zpz(DisasContext *s, int zt, int pg, int zm, tcg_temp_free_i32(t_desc); } -/* Indexed by [be][ff][xs][u][msz]. */ -static gen_helper_gvec_mem_scatter * const gather_load_fn32[2][2][2][2][3] = { - /* Little-endian */ - { { { { gen_helper_sve_ldbss_zsu, - gen_helper_sve_ldhss_le_zsu, - NULL, }, - { gen_helper_sve_ldbsu_zsu, - gen_helper_sve_ldhsu_le_zsu, - gen_helper_sve_ldss_le_zsu, } }, - { { gen_helper_sve_ldbss_zss, - gen_helper_sve_ldhss_le_zss, - NULL, }, - { gen_helper_sve_ldbsu_zss, - gen_helper_sve_ldhsu_le_zss, - gen_helper_sve_ldss_le_zss, } } }, +/* Indexed by [mte][be][ff][xs][u][msz]. */ +static gen_helper_gvec_mem_scatter * const +gather_load_fn32[2][2][2][2][2][3] = { + { /* MTE Inactive */ + { /* Little-endian */ + { { { gen_helper_sve_ldbss_zsu, + gen_helper_sve_ldhss_le_zsu, + NULL, }, + { gen_helper_sve_ldbsu_zsu, + gen_helper_sve_ldhsu_le_zsu, + gen_helper_sve_ldss_le_zsu, } }, + { { gen_helper_sve_ldbss_zss, + gen_helper_sve_ldhss_le_zss, + NULL, }, + { gen_helper_sve_ldbsu_zss, + gen_helper_sve_ldhsu_le_zss, + gen_helper_sve_ldss_le_zss, } } }, - /* First-fault */ - { { { gen_helper_sve_ldffbss_zsu, - gen_helper_sve_ldffhss_le_zsu, - NULL, }, - { gen_helper_sve_ldffbsu_zsu, - gen_helper_sve_ldffhsu_le_zsu, - gen_helper_sve_ldffss_le_zsu, } }, - { { gen_helper_sve_ldffbss_zss, - gen_helper_sve_ldffhss_le_zss, - NULL, }, - { gen_helper_sve_ldffbsu_zss, - gen_helper_sve_ldffhsu_le_zss, - gen_helper_sve_ldffss_le_zss, } } } }, + /* First-fault */ + { { { gen_helper_sve_ldffbss_zsu, + gen_helper_sve_ldffhss_le_zsu, + NULL, }, + { gen_helper_sve_ldffbsu_zsu, + gen_helper_sve_ldffhsu_le_zsu, + gen_helper_sve_ldffss_le_zsu, } }, + { { gen_helper_sve_ldffbss_zss, + gen_helper_sve_ldffhss_le_zss, + NULL, }, + { gen_helper_sve_ldffbsu_zss, + gen_helper_sve_ldffhsu_le_zss, + gen_helper_sve_ldffss_le_zss, } } } }, - /* Big-endian */ - { { { { gen_helper_sve_ldbss_zsu, - gen_helper_sve_ldhss_be_zsu, - NULL, }, - { gen_helper_sve_ldbsu_zsu, - gen_helper_sve_ldhsu_be_zsu, - gen_helper_sve_ldss_be_zsu, } }, - { { gen_helper_sve_ldbss_zss, - gen_helper_sve_ldhss_be_zss, - NULL, }, - { gen_helper_sve_ldbsu_zss, - gen_helper_sve_ldhsu_be_zss, - gen_helper_sve_ldss_be_zss, } } }, + { /* Big-endian */ + { { { gen_helper_sve_ldbss_zsu, + gen_helper_sve_ldhss_be_zsu, + NULL, }, + { gen_helper_sve_ldbsu_zsu, + gen_helper_sve_ldhsu_be_zsu, + gen_helper_sve_ldss_be_zsu, } }, + { { gen_helper_sve_ldbss_zss, + gen_helper_sve_ldhss_be_zss, + NULL, }, + { gen_helper_sve_ldbsu_zss, + gen_helper_sve_ldhsu_be_zss, + gen_helper_sve_ldss_be_zss, } } }, - /* First-fault */ - { { { gen_helper_sve_ldffbss_zsu, - gen_helper_sve_ldffhss_be_zsu, - NULL, }, - { gen_helper_sve_ldffbsu_zsu, - gen_helper_sve_ldffhsu_be_zsu, - gen_helper_sve_ldffss_be_zsu, } }, - { { gen_helper_sve_ldffbss_zss, - gen_helper_sve_ldffhss_be_zss, - NULL, }, - { gen_helper_sve_ldffbsu_zss, - gen_helper_sve_ldffhsu_be_zss, - gen_helper_sve_ldffss_be_zss, } } } }, + /* First-fault */ + { { { gen_helper_sve_ldffbss_zsu, + gen_helper_sve_ldffhss_be_zsu, + NULL, }, + { gen_helper_sve_ldffbsu_zsu, + gen_helper_sve_ldffhsu_be_zsu, + gen_helper_sve_ldffss_be_zsu, } }, + { { gen_helper_sve_ldffbss_zss, + gen_helper_sve_ldffhss_be_zss, + NULL, }, + { gen_helper_sve_ldffbsu_zss, + gen_helper_sve_ldffhsu_be_zss, + gen_helper_sve_ldffss_be_zss, } } } } }, + { /* MTE Active */ + { /* Little-endian */ + { { { gen_helper_sve_ldbss_zsu_mte, + gen_helper_sve_ldhss_le_zsu_mte, + NULL, }, + { gen_helper_sve_ldbsu_zsu_mte, + gen_helper_sve_ldhsu_le_zsu_mte, + gen_helper_sve_ldss_le_zsu_mte, } }, + { { gen_helper_sve_ldbss_zss_mte, + gen_helper_sve_ldhss_le_zss_mte, + NULL, }, + { gen_helper_sve_ldbsu_zss_mte, + gen_helper_sve_ldhsu_le_zss_mte, + gen_helper_sve_ldss_le_zss_mte, } } }, + + /* First-fault */ + { { { gen_helper_sve_ldffbss_zsu_mte, + gen_helper_sve_ldffhss_le_zsu_mte, + NULL, }, + { gen_helper_sve_ldffbsu_zsu_mte, + gen_helper_sve_ldffhsu_le_zsu_mte, + gen_helper_sve_ldffss_le_zsu_mte, } }, + { { gen_helper_sve_ldffbss_zss_mte, + gen_helper_sve_ldffhss_le_zss_mte, + NULL, }, + { gen_helper_sve_ldffbsu_zss_mte, + gen_helper_sve_ldffhsu_le_zss_mte, + gen_helper_sve_ldffss_le_zss_mte, } } } }, + + { /* Big-endian */ + { { { gen_helper_sve_ldbss_zsu_mte, + gen_helper_sve_ldhss_be_zsu_mte, + NULL, }, + { gen_helper_sve_ldbsu_zsu_mte, + gen_helper_sve_ldhsu_be_zsu_mte, + gen_helper_sve_ldss_be_zsu_mte, } }, + { { gen_helper_sve_ldbss_zss_mte, + gen_helper_sve_ldhss_be_zss_mte, + NULL, }, + { gen_helper_sve_ldbsu_zss_mte, + gen_helper_sve_ldhsu_be_zss_mte, + gen_helper_sve_ldss_be_zss_mte, } } }, + + /* First-fault */ + { { { gen_helper_sve_ldffbss_zsu_mte, + gen_helper_sve_ldffhss_be_zsu_mte, + NULL, }, + { gen_helper_sve_ldffbsu_zsu_mte, + gen_helper_sve_ldffhsu_be_zsu_mte, + gen_helper_sve_ldffss_be_zsu_mte, } }, + { { gen_helper_sve_ldffbss_zss_mte, + gen_helper_sve_ldffhss_be_zss_mte, + NULL, }, + { gen_helper_sve_ldffbsu_zss_mte, + gen_helper_sve_ldffhsu_be_zss_mte, + gen_helper_sve_ldffss_be_zss_mte, } } } } }, }; /* Note that we overload xs=2 to indicate 64-bit offset. */ -static gen_helper_gvec_mem_scatter * const gather_load_fn64[2][2][3][2][4] = { - /* Little-endian */ - { { { { gen_helper_sve_ldbds_zsu, - gen_helper_sve_ldhds_le_zsu, - gen_helper_sve_ldsds_le_zsu, - NULL, }, - { gen_helper_sve_ldbdu_zsu, - gen_helper_sve_ldhdu_le_zsu, - gen_helper_sve_ldsdu_le_zsu, - gen_helper_sve_lddd_le_zsu, } }, - { { gen_helper_sve_ldbds_zss, - gen_helper_sve_ldhds_le_zss, - gen_helper_sve_ldsds_le_zss, - NULL, }, - { gen_helper_sve_ldbdu_zss, - gen_helper_sve_ldhdu_le_zss, - gen_helper_sve_ldsdu_le_zss, - gen_helper_sve_lddd_le_zss, } }, - { { gen_helper_sve_ldbds_zd, - gen_helper_sve_ldhds_le_zd, - gen_helper_sve_ldsds_le_zd, - NULL, }, - { gen_helper_sve_ldbdu_zd, - gen_helper_sve_ldhdu_le_zd, - gen_helper_sve_ldsdu_le_zd, - gen_helper_sve_lddd_le_zd, } } }, +static gen_helper_gvec_mem_scatter * const +gather_load_fn64[2][2][2][3][2][4] = { + { /* MTE Inactive */ + { /* Little-endian */ + { { { gen_helper_sve_ldbds_zsu, + gen_helper_sve_ldhds_le_zsu, + gen_helper_sve_ldsds_le_zsu, + NULL, }, + { gen_helper_sve_ldbdu_zsu, + gen_helper_sve_ldhdu_le_zsu, + gen_helper_sve_ldsdu_le_zsu, + gen_helper_sve_lddd_le_zsu, } }, + { { gen_helper_sve_ldbds_zss, + gen_helper_sve_ldhds_le_zss, + gen_helper_sve_ldsds_le_zss, + NULL, }, + { gen_helper_sve_ldbdu_zss, + gen_helper_sve_ldhdu_le_zss, + gen_helper_sve_ldsdu_le_zss, + gen_helper_sve_lddd_le_zss, } }, + { { gen_helper_sve_ldbds_zd, + gen_helper_sve_ldhds_le_zd, + gen_helper_sve_ldsds_le_zd, + NULL, }, + { gen_helper_sve_ldbdu_zd, + gen_helper_sve_ldhdu_le_zd, + gen_helper_sve_ldsdu_le_zd, + gen_helper_sve_lddd_le_zd, } } }, - /* First-fault */ - { { { gen_helper_sve_ldffbds_zsu, - gen_helper_sve_ldffhds_le_zsu, - gen_helper_sve_ldffsds_le_zsu, - NULL, }, - { gen_helper_sve_ldffbdu_zsu, - gen_helper_sve_ldffhdu_le_zsu, - gen_helper_sve_ldffsdu_le_zsu, - gen_helper_sve_ldffdd_le_zsu, } }, - { { gen_helper_sve_ldffbds_zss, - gen_helper_sve_ldffhds_le_zss, - gen_helper_sve_ldffsds_le_zss, - NULL, }, - { gen_helper_sve_ldffbdu_zss, - gen_helper_sve_ldffhdu_le_zss, - gen_helper_sve_ldffsdu_le_zss, - gen_helper_sve_ldffdd_le_zss, } }, - { { gen_helper_sve_ldffbds_zd, - gen_helper_sve_ldffhds_le_zd, - gen_helper_sve_ldffsds_le_zd, - NULL, }, - { gen_helper_sve_ldffbdu_zd, - gen_helper_sve_ldffhdu_le_zd, - gen_helper_sve_ldffsdu_le_zd, - gen_helper_sve_ldffdd_le_zd, } } } }, + /* First-fault */ + { { { gen_helper_sve_ldffbds_zsu, + gen_helper_sve_ldffhds_le_zsu, + gen_helper_sve_ldffsds_le_zsu, + NULL, }, + { gen_helper_sve_ldffbdu_zsu, + gen_helper_sve_ldffhdu_le_zsu, + gen_helper_sve_ldffsdu_le_zsu, + gen_helper_sve_ldffdd_le_zsu, } }, + { { gen_helper_sve_ldffbds_zss, + gen_helper_sve_ldffhds_le_zss, + gen_helper_sve_ldffsds_le_zss, + NULL, }, + { gen_helper_sve_ldffbdu_zss, + gen_helper_sve_ldffhdu_le_zss, + gen_helper_sve_ldffsdu_le_zss, + gen_helper_sve_ldffdd_le_zss, } }, + { { gen_helper_sve_ldffbds_zd, + gen_helper_sve_ldffhds_le_zd, + gen_helper_sve_ldffsds_le_zd, + NULL, }, + { gen_helper_sve_ldffbdu_zd, + gen_helper_sve_ldffhdu_le_zd, + gen_helper_sve_ldffsdu_le_zd, + gen_helper_sve_ldffdd_le_zd, } } } }, + { /* Big-endian */ + { { { gen_helper_sve_ldbds_zsu, + gen_helper_sve_ldhds_be_zsu, + gen_helper_sve_ldsds_be_zsu, + NULL, }, + { gen_helper_sve_ldbdu_zsu, + gen_helper_sve_ldhdu_be_zsu, + gen_helper_sve_ldsdu_be_zsu, + gen_helper_sve_lddd_be_zsu, } }, + { { gen_helper_sve_ldbds_zss, + gen_helper_sve_ldhds_be_zss, + gen_helper_sve_ldsds_be_zss, + NULL, }, + { gen_helper_sve_ldbdu_zss, + gen_helper_sve_ldhdu_be_zss, + gen_helper_sve_ldsdu_be_zss, + gen_helper_sve_lddd_be_zss, } }, + { { gen_helper_sve_ldbds_zd, + gen_helper_sve_ldhds_be_zd, + gen_helper_sve_ldsds_be_zd, + NULL, }, + { gen_helper_sve_ldbdu_zd, + gen_helper_sve_ldhdu_be_zd, + gen_helper_sve_ldsdu_be_zd, + gen_helper_sve_lddd_be_zd, } } }, - /* Big-endian */ - { { { { gen_helper_sve_ldbds_zsu, - gen_helper_sve_ldhds_be_zsu, - gen_helper_sve_ldsds_be_zsu, - NULL, }, - { gen_helper_sve_ldbdu_zsu, - gen_helper_sve_ldhdu_be_zsu, - gen_helper_sve_ldsdu_be_zsu, - gen_helper_sve_lddd_be_zsu, } }, - { { gen_helper_sve_ldbds_zss, - gen_helper_sve_ldhds_be_zss, - gen_helper_sve_ldsds_be_zss, - NULL, }, - { gen_helper_sve_ldbdu_zss, - gen_helper_sve_ldhdu_be_zss, - gen_helper_sve_ldsdu_be_zss, - gen_helper_sve_lddd_be_zss, } }, - { { gen_helper_sve_ldbds_zd, - gen_helper_sve_ldhds_be_zd, - gen_helper_sve_ldsds_be_zd, - NULL, }, - { gen_helper_sve_ldbdu_zd, - gen_helper_sve_ldhdu_be_zd, - gen_helper_sve_ldsdu_be_zd, - gen_helper_sve_lddd_be_zd, } } }, + /* First-fault */ + { { { gen_helper_sve_ldffbds_zsu, + gen_helper_sve_ldffhds_be_zsu, + gen_helper_sve_ldffsds_be_zsu, + NULL, }, + { gen_helper_sve_ldffbdu_zsu, + gen_helper_sve_ldffhdu_be_zsu, + gen_helper_sve_ldffsdu_be_zsu, + gen_helper_sve_ldffdd_be_zsu, } }, + { { gen_helper_sve_ldffbds_zss, + gen_helper_sve_ldffhds_be_zss, + gen_helper_sve_ldffsds_be_zss, + NULL, }, + { gen_helper_sve_ldffbdu_zss, + gen_helper_sve_ldffhdu_be_zss, + gen_helper_sve_ldffsdu_be_zss, + gen_helper_sve_ldffdd_be_zss, } }, + { { gen_helper_sve_ldffbds_zd, + gen_helper_sve_ldffhds_be_zd, + gen_helper_sve_ldffsds_be_zd, + NULL, }, + { gen_helper_sve_ldffbdu_zd, + gen_helper_sve_ldffhdu_be_zd, + gen_helper_sve_ldffsdu_be_zd, + gen_helper_sve_ldffdd_be_zd, } } } } }, + { /* MTE Active */ + { /* Little-endian */ + { { { gen_helper_sve_ldbds_zsu_mte, + gen_helper_sve_ldhds_le_zsu_mte, + gen_helper_sve_ldsds_le_zsu_mte, + NULL, }, + { gen_helper_sve_ldbdu_zsu_mte, + gen_helper_sve_ldhdu_le_zsu_mte, + gen_helper_sve_ldsdu_le_zsu_mte, + gen_helper_sve_lddd_le_zsu_mte, } }, + { { gen_helper_sve_ldbds_zss_mte, + gen_helper_sve_ldhds_le_zss_mte, + gen_helper_sve_ldsds_le_zss_mte, + NULL, }, + { gen_helper_sve_ldbdu_zss_mte, + gen_helper_sve_ldhdu_le_zss_mte, + gen_helper_sve_ldsdu_le_zss_mte, + gen_helper_sve_lddd_le_zss_mte, } }, + { { gen_helper_sve_ldbds_zd_mte, + gen_helper_sve_ldhds_le_zd_mte, + gen_helper_sve_ldsds_le_zd_mte, + NULL, }, + { gen_helper_sve_ldbdu_zd_mte, + gen_helper_sve_ldhdu_le_zd_mte, + gen_helper_sve_ldsdu_le_zd_mte, + gen_helper_sve_lddd_le_zd_mte, } } }, - /* First-fault */ - { { { gen_helper_sve_ldffbds_zsu, - gen_helper_sve_ldffhds_be_zsu, - gen_helper_sve_ldffsds_be_zsu, - NULL, }, - { gen_helper_sve_ldffbdu_zsu, - gen_helper_sve_ldffhdu_be_zsu, - gen_helper_sve_ldffsdu_be_zsu, - gen_helper_sve_ldffdd_be_zsu, } }, - { { gen_helper_sve_ldffbds_zss, - gen_helper_sve_ldffhds_be_zss, - gen_helper_sve_ldffsds_be_zss, - NULL, }, - { gen_helper_sve_ldffbdu_zss, - gen_helper_sve_ldffhdu_be_zss, - gen_helper_sve_ldffsdu_be_zss, - gen_helper_sve_ldffdd_be_zss, } }, - { { gen_helper_sve_ldffbds_zd, - gen_helper_sve_ldffhds_be_zd, - gen_helper_sve_ldffsds_be_zd, - NULL, }, - { gen_helper_sve_ldffbdu_zd, - gen_helper_sve_ldffhdu_be_zd, - gen_helper_sve_ldffsdu_be_zd, - gen_helper_sve_ldffdd_be_zd, } } } }, + /* First-fault */ + { { { gen_helper_sve_ldffbds_zsu_mte, + gen_helper_sve_ldffhds_le_zsu_mte, + gen_helper_sve_ldffsds_le_zsu_mte, + NULL, }, + { gen_helper_sve_ldffbdu_zsu_mte, + gen_helper_sve_ldffhdu_le_zsu_mte, + gen_helper_sve_ldffsdu_le_zsu_mte, + gen_helper_sve_ldffdd_le_zsu_mte, } }, + { { gen_helper_sve_ldffbds_zss_mte, + gen_helper_sve_ldffhds_le_zss_mte, + gen_helper_sve_ldffsds_le_zss_mte, + NULL, }, + { gen_helper_sve_ldffbdu_zss_mte, + gen_helper_sve_ldffhdu_le_zss_mte, + gen_helper_sve_ldffsdu_le_zss_mte, + gen_helper_sve_ldffdd_le_zss_mte, } }, + { { gen_helper_sve_ldffbds_zd_mte, + gen_helper_sve_ldffhds_le_zd_mte, + gen_helper_sve_ldffsds_le_zd_mte, + NULL, }, + { gen_helper_sve_ldffbdu_zd_mte, + gen_helper_sve_ldffhdu_le_zd_mte, + gen_helper_sve_ldffsdu_le_zd_mte, + gen_helper_sve_ldffdd_le_zd_mte, } } } }, + { /* Big-endian */ + { { { gen_helper_sve_ldbds_zsu_mte, + gen_helper_sve_ldhds_be_zsu_mte, + gen_helper_sve_ldsds_be_zsu_mte, + NULL, }, + { gen_helper_sve_ldbdu_zsu_mte, + gen_helper_sve_ldhdu_be_zsu_mte, + gen_helper_sve_ldsdu_be_zsu_mte, + gen_helper_sve_lddd_be_zsu_mte, } }, + { { gen_helper_sve_ldbds_zss_mte, + gen_helper_sve_ldhds_be_zss_mte, + gen_helper_sve_ldsds_be_zss_mte, + NULL, }, + { gen_helper_sve_ldbdu_zss_mte, + gen_helper_sve_ldhdu_be_zss_mte, + gen_helper_sve_ldsdu_be_zss_mte, + gen_helper_sve_lddd_be_zss_mte, } }, + { { gen_helper_sve_ldbds_zd_mte, + gen_helper_sve_ldhds_be_zd_mte, + gen_helper_sve_ldsds_be_zd_mte, + NULL, }, + { gen_helper_sve_ldbdu_zd_mte, + gen_helper_sve_ldhdu_be_zd_mte, + gen_helper_sve_ldsdu_be_zd_mte, + gen_helper_sve_lddd_be_zd_mte, } } }, + + /* First-fault */ + { { { gen_helper_sve_ldffbds_zsu_mte, + gen_helper_sve_ldffhds_be_zsu_mte, + gen_helper_sve_ldffsds_be_zsu_mte, + NULL, }, + { gen_helper_sve_ldffbdu_zsu_mte, + gen_helper_sve_ldffhdu_be_zsu_mte, + gen_helper_sve_ldffsdu_be_zsu_mte, + gen_helper_sve_ldffdd_be_zsu_mte, } }, + { { gen_helper_sve_ldffbds_zss_mte, + gen_helper_sve_ldffhds_be_zss_mte, + gen_helper_sve_ldffsds_be_zss_mte, + NULL, }, + { gen_helper_sve_ldffbdu_zss_mte, + gen_helper_sve_ldffhdu_be_zss_mte, + gen_helper_sve_ldffsdu_be_zss_mte, + gen_helper_sve_ldffdd_be_zss_mte, } }, + { { gen_helper_sve_ldffbds_zd_mte, + gen_helper_sve_ldffhds_be_zd_mte, + gen_helper_sve_ldffsds_be_zd_mte, + NULL, }, + { gen_helper_sve_ldffbdu_zd_mte, + gen_helper_sve_ldffhdu_be_zd_mte, + gen_helper_sve_ldffsdu_be_zd_mte, + gen_helper_sve_ldffdd_be_zd_mte, } } } } }, }; static bool trans_LD1_zprz(DisasContext *s, arg_LD1_zprz *a) { gen_helper_gvec_mem_scatter *fn = NULL; - int be = s->be_data == MO_BE; + bool be = s->be_data == MO_BE; + bool mte = s->mte_active[0]; if (!sve_access_check(s)) { return true; @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zprz(DisasContext *s, arg_LD1_zprz *a) switch (a->esz) { case MO_32: - fn = gather_load_fn32[be][a->ff][a->xs][a->u][a->msz]; + fn = gather_load_fn32[mte][be][a->ff][a->xs][a->u][a->msz]; break; case MO_64: - fn = gather_load_fn64[be][a->ff][a->xs][a->u][a->msz]; + fn = gather_load_fn64[mte][be][a->ff][a->xs][a->u][a->msz]; break; } assert(fn != NULL); do_mem_zpz(s, a->rd, a->pg, a->rm, a->scale * a->msz, - cpu_reg_sp(s, a->rn), a->msz, fn); + cpu_reg_sp(s, a->rn), a->msz, false, fn); return true; } static bool trans_LD1_zpiz(DisasContext *s, arg_LD1_zpiz *a) { gen_helper_gvec_mem_scatter *fn = NULL; - int be = s->be_data == MO_BE; + bool be = s->be_data == MO_BE; + bool mte = s->mte_active[0]; TCGv_i64 imm; if (a->esz < a->msz || (a->esz == a->msz && !a->u)) { @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zpiz(DisasContext *s, arg_LD1_zpiz *a) switch (a->esz) { case MO_32: - fn = gather_load_fn32[be][a->ff][0][a->u][a->msz]; + fn = gather_load_fn32[mte][be][a->ff][0][a->u][a->msz]; break; case MO_64: - fn = gather_load_fn64[be][a->ff][2][a->u][a->msz]; + fn = gather_load_fn64[mte][be][a->ff][2][a->u][a->msz]; break; } assert(fn != NULL); @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zpiz(DisasContext *s, arg_LD1_zpiz *a) * by loading the immediate into the scalar parameter. */ imm = tcg_const_i64(a->imm << a->msz); - do_mem_zpz(s, a->rd, a->pg, a->rn, 0, imm, a->msz, fn); + do_mem_zpz(s, a->rd, a->pg, a->rn, 0, imm, a->msz, false, fn); tcg_temp_free_i64(imm); return true; } -/* Indexed by [be][xs][msz]. */ -static gen_helper_gvec_mem_scatter * const scatter_store_fn32[2][2][3] = { - /* Little-endian */ - { { gen_helper_sve_stbs_zsu, - gen_helper_sve_sths_le_zsu, - gen_helper_sve_stss_le_zsu, }, - { gen_helper_sve_stbs_zss, - gen_helper_sve_sths_le_zss, - gen_helper_sve_stss_le_zss, } }, - /* Big-endian */ - { { gen_helper_sve_stbs_zsu, - gen_helper_sve_sths_be_zsu, - gen_helper_sve_stss_be_zsu, }, - { gen_helper_sve_stbs_zss, - gen_helper_sve_sths_be_zss, - gen_helper_sve_stss_be_zss, } }, +/* Indexed by [mte][be][xs][msz]. */ +static gen_helper_gvec_mem_scatter * const scatter_store_fn32[2][2][2][3] = { + { /* MTE Inactive */ + { /* Little-endian */ + { gen_helper_sve_stbs_zsu, + gen_helper_sve_sths_le_zsu, + gen_helper_sve_stss_le_zsu, }, + { gen_helper_sve_stbs_zss, + gen_helper_sve_sths_le_zss, + gen_helper_sve_stss_le_zss, } }, + { /* Big-endian */ + { gen_helper_sve_stbs_zsu, + gen_helper_sve_sths_be_zsu, + gen_helper_sve_stss_be_zsu, }, + { gen_helper_sve_stbs_zss, + gen_helper_sve_sths_be_zss, + gen_helper_sve_stss_be_zss, } } }, + { /* MTE Active */ + { /* Little-endian */ + { gen_helper_sve_stbs_zsu_mte, + gen_helper_sve_sths_le_zsu_mte, + gen_helper_sve_stss_le_zsu_mte, }, + { gen_helper_sve_stbs_zss_mte, + gen_helper_sve_sths_le_zss_mte, + gen_helper_sve_stss_le_zss_mte, } }, + { /* Big-endian */ + { gen_helper_sve_stbs_zsu_mte, + gen_helper_sve_sths_be_zsu_mte, + gen_helper_sve_stss_be_zsu_mte, }, + { gen_helper_sve_stbs_zss_mte, + gen_helper_sve_sths_be_zss_mte, + gen_helper_sve_stss_be_zss_mte, } } }, }; /* Note that we overload xs=2 to indicate 64-bit offset. */ -static gen_helper_gvec_mem_scatter * const scatter_store_fn64[2][3][4] = { - /* Little-endian */ - { { gen_helper_sve_stbd_zsu, - gen_helper_sve_sthd_le_zsu, - gen_helper_sve_stsd_le_zsu, - gen_helper_sve_stdd_le_zsu, }, - { gen_helper_sve_stbd_zss, - gen_helper_sve_sthd_le_zss, - gen_helper_sve_stsd_le_zss, - gen_helper_sve_stdd_le_zss, }, - { gen_helper_sve_stbd_zd, - gen_helper_sve_sthd_le_zd, - gen_helper_sve_stsd_le_zd, - gen_helper_sve_stdd_le_zd, } }, - /* Big-endian */ - { { gen_helper_sve_stbd_zsu, - gen_helper_sve_sthd_be_zsu, - gen_helper_sve_stsd_be_zsu, - gen_helper_sve_stdd_be_zsu, }, - { gen_helper_sve_stbd_zss, - gen_helper_sve_sthd_be_zss, - gen_helper_sve_stsd_be_zss, - gen_helper_sve_stdd_be_zss, }, - { gen_helper_sve_stbd_zd, - gen_helper_sve_sthd_be_zd, - gen_helper_sve_stsd_be_zd, - gen_helper_sve_stdd_be_zd, } }, +static gen_helper_gvec_mem_scatter * const scatter_store_fn64[2][2][3][4] = { + { /* MTE Inactive */ + { /* Little-endian */ + { gen_helper_sve_stbd_zsu, + gen_helper_sve_sthd_le_zsu, + gen_helper_sve_stsd_le_zsu, + gen_helper_sve_stdd_le_zsu, }, + { gen_helper_sve_stbd_zss, + gen_helper_sve_sthd_le_zss, + gen_helper_sve_stsd_le_zss, + gen_helper_sve_stdd_le_zss, }, + { gen_helper_sve_stbd_zd, + gen_helper_sve_sthd_le_zd, + gen_helper_sve_stsd_le_zd, + gen_helper_sve_stdd_le_zd, } }, + { /* Big-endian */ + { gen_helper_sve_stbd_zsu, + gen_helper_sve_sthd_be_zsu, + gen_helper_sve_stsd_be_zsu, + gen_helper_sve_stdd_be_zsu, }, + { gen_helper_sve_stbd_zss, + gen_helper_sve_sthd_be_zss, + gen_helper_sve_stsd_be_zss, + gen_helper_sve_stdd_be_zss, }, + { gen_helper_sve_stbd_zd, + gen_helper_sve_sthd_be_zd, + gen_helper_sve_stsd_be_zd, + gen_helper_sve_stdd_be_zd, } } }, + { /* MTE Inactive */ + { /* Little-endian */ + { gen_helper_sve_stbd_zsu_mte, + gen_helper_sve_sthd_le_zsu_mte, + gen_helper_sve_stsd_le_zsu_mte, + gen_helper_sve_stdd_le_zsu_mte, }, + { gen_helper_sve_stbd_zss_mte, + gen_helper_sve_sthd_le_zss_mte, + gen_helper_sve_stsd_le_zss_mte, + gen_helper_sve_stdd_le_zss_mte, }, + { gen_helper_sve_stbd_zd_mte, + gen_helper_sve_sthd_le_zd_mte, + gen_helper_sve_stsd_le_zd_mte, + gen_helper_sve_stdd_le_zd_mte, } }, + { /* Big-endian */ + { gen_helper_sve_stbd_zsu_mte, + gen_helper_sve_sthd_be_zsu_mte, + gen_helper_sve_stsd_be_zsu_mte, + gen_helper_sve_stdd_be_zsu_mte, }, + { gen_helper_sve_stbd_zss_mte, + gen_helper_sve_sthd_be_zss_mte, + gen_helper_sve_stsd_be_zss_mte, + gen_helper_sve_stdd_be_zss_mte, }, + { gen_helper_sve_stbd_zd_mte, + gen_helper_sve_sthd_be_zd_mte, + gen_helper_sve_stsd_be_zd_mte, + gen_helper_sve_stdd_be_zd_mte, } } }, }; static bool trans_ST1_zprz(DisasContext *s, arg_ST1_zprz *a) { gen_helper_gvec_mem_scatter *fn; - int be = s->be_data == MO_BE; + bool be = s->be_data == MO_BE; + bool mte = s->mte_active[0]; if (a->esz < a->msz || (a->msz == 0 && a->scale)) { return false; @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zprz(DisasContext *s, arg_ST1_zprz *a) } switch (a->esz) { case MO_32: - fn = scatter_store_fn32[be][a->xs][a->msz]; + fn = scatter_store_fn32[mte][be][a->xs][a->msz]; break; case MO_64: - fn = scatter_store_fn64[be][a->xs][a->msz]; + fn = scatter_store_fn64[mte][be][a->xs][a->msz]; break; default: g_assert_not_reached(); } do_mem_zpz(s, a->rd, a->pg, a->rm, a->scale * a->msz, - cpu_reg_sp(s, a->rn), a->msz, fn); + cpu_reg_sp(s, a->rn), a->msz, true, fn); return true; } static bool trans_ST1_zpiz(DisasContext *s, arg_ST1_zpiz *a) { gen_helper_gvec_mem_scatter *fn = NULL; - int be = s->be_data == MO_BE; + bool be = s->be_data == MO_BE; + bool mte = s->mte_active[0]; TCGv_i64 imm; if (a->esz < a->msz) { @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zpiz(DisasContext *s, arg_ST1_zpiz *a) switch (a->esz) { case MO_32: - fn = scatter_store_fn32[be][0][a->msz]; + fn = scatter_store_fn32[mte][be][0][a->msz]; break; case MO_64: - fn = scatter_store_fn64[be][2][a->msz]; + fn = scatter_store_fn64[mte][be][2][a->msz]; break; } assert(fn != NULL); @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zpiz(DisasContext *s, arg_ST1_zpiz *a) * by loading the immediate into the scalar parameter. */ imm = tcg_const_i64(a->imm << a->msz); - do_mem_zpz(s, a->rd, a->pg, a->rn, 0, imm, a->msz, fn); + do_mem_zpz(s, a->rd, a->pg, a->rn, 0, imm, a->msz, true, fn); tcg_temp_free_i64(imm); return true; } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> There are a number of paths by which the TBI is still intact for user-only in the SVE helpers. Because we currently always set TBI for user-only, we do not need to pass down the actual TBI setting from above, and we can remove the top byte in the inner-most primitives, so that none are forgotten. Moreover, this keeps the "dirty" pointer around at the higher levels, where we need it for any MTE checking. Since the normal case, especially for user-only, goes through RAM, this clearing merely adds two insns per page lookup, which will be completely in the noise. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-39-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 3 +++ target/arm/sve_helper.c | 19 +++++++++++++++++-- target/arm/translate-a64.c | 5 +++++ 3 files changed, 25 insertions(+), 2 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) * Enable TBI0 and TBI1. While the real kernel only enables TBI0, * turning on both here will produce smaller code and otherwise * make no difference to the user-level emulation. + * + * In sve_probe_page, we assume that this is set. + * Do not modify this without other changes. */ env->cp15.tcr_el[1].raw_tcr = (3ULL << 37); #else diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ static void sve_##NAME##_host(void *vd, intptr_t reg_off, void *host) \ static void sve_##NAME##_tlb(CPUARMState *env, void *vd, intptr_t reg_off, \ target_ulong addr, uintptr_t ra) \ { \ - *(TYPEE *)(vd + H(reg_off)) = (TYPEM)TLB(env, addr, ra); \ + *(TYPEE *)(vd + H(reg_off)) = \ + (TYPEM)TLB(env, useronly_clean_ptr(addr), ra); \ } #define DO_ST_TLB(NAME, H, TYPEE, TYPEM, TLB) \ static void sve_##NAME##_tlb(CPUARMState *env, void *vd, intptr_t reg_off, \ target_ulong addr, uintptr_t ra) \ { \ - TLB(env, addr, (TYPEM)*(TYPEE *)(vd + H(reg_off)), ra); \ + TLB(env, useronly_clean_ptr(addr), \ + (TYPEM)*(TYPEE *)(vd + H(reg_off)), ra); \ } #define DO_LD_PRIM_1(NAME, H, TE, TM) \ @@ -XXX,XX +XXX,XX @@ static bool sve_probe_page(SVEHostPage *info, bool nofault, int flags; addr += mem_off; + + /* + * User-only currently always issues with TBI. See the comment + * above useronly_clean_ptr. Usually we clean this top byte away + * during translation, but we can't do that for e.g. vector + imm + * addressing modes. + * + * We currently always enable TBI for user-only, and do not provide + * a way to turn it off. So clean the pointer unconditionally here, + * rather than look it up here, or pass it down from above. + */ + addr = useronly_clean_ptr(addr); + flags = probe_access_flags(env, addr, access_type, mmu_idx, nofault, &info->host, retaddr); info->flags = flags; 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_init_disas_context(DisasContextBase *dcbase, dc->features = env->features; dc->dcz_blocksize = arm_cpu->dcz_blocksize; +#ifdef CONFIG_USER_ONLY + /* In sve_probe_page, we assume TBI is enabled. */ + tcg_debug_assert(dc->tbid & 1); +#endif + /* Single step state. The code-generation logic here is: * SS_ACTIVE == 0: * generate code with no special handling for single-stepping (except -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> This is DC GVA and DC GZVA, and the tag check for DC ZVA. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-40-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 4 +++- target/arm/helper.c | 16 ++++++++++++++++ target/arm/translate-a64.c | 39 ++++++++++++++++++++++++++++++++++++++ 3 files changed, 58 insertions(+), 1 deletion(-) 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 uint64_t cpreg_to_kvm_id(uint32_t cpregid) #define ARM_CP_NZCV (ARM_CP_SPECIAL | 0x0300) #define ARM_CP_CURRENTEL (ARM_CP_SPECIAL | 0x0400) #define ARM_CP_DC_ZVA (ARM_CP_SPECIAL | 0x0500) -#define ARM_LAST_SPECIAL ARM_CP_DC_ZVA +#define ARM_CP_DC_GVA (ARM_CP_SPECIAL | 0x0600) +#define ARM_CP_DC_GZVA (ARM_CP_SPECIAL | 0x0700) +#define ARM_LAST_SPECIAL ARM_CP_DC_GZVA #define ARM_CP_FPU 0x1000 #define ARM_CP_SVE 0x2000 #define ARM_CP_NO_GDB 0x4000 diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo mte_el0_cacheop_reginfo[] = { .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 14, .opc2 = 5, .type = ARM_CP_NOP, .access = PL0_W, .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_GVA", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 4, .opc2 = 3, + .access = PL0_W, .type = ARM_CP_DC_GVA, +#ifndef CONFIG_USER_ONLY + /* Avoid overhead of an access check that always passes in user-mode */ + .accessfn = aa64_zva_access, +#endif + }, + { .name = "DC_GZVA", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 4, .opc2 = 4, + .access = PL0_W, .type = ARM_CP_DC_GZVA, +#ifndef CONFIG_USER_ONLY + /* Avoid overhead of an access check that always passes in user-mode */ + .accessfn = aa64_zva_access, +#endif + }, REGINFO_SENTINEL }; diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread, } gen_helper_dc_zva(cpu_env, tcg_rt); return; + case ARM_CP_DC_GVA: + { + TCGv_i64 clean_addr, tag; + + /* + * DC_GVA, like DC_ZVA, requires that we supply the original + * pointer for an invalid page. Probe that address first. + */ + tcg_rt = cpu_reg(s, rt); + clean_addr = clean_data_tbi(s, tcg_rt); + gen_probe_access(s, clean_addr, MMU_DATA_STORE, MO_8); + + if (s->ata) { + /* Extract the tag from the register to match STZGM. */ + tag = tcg_temp_new_i64(); + tcg_gen_shri_i64(tag, tcg_rt, 56); + gen_helper_stzgm_tags(cpu_env, clean_addr, tag); + tcg_temp_free_i64(tag); + } + } + return; + case ARM_CP_DC_GZVA: + { + TCGv_i64 clean_addr, tag; + + /* For DC_GZVA, we can rely on DC_ZVA for the proper fault. */ + tcg_rt = cpu_reg(s, rt); + clean_addr = clean_data_tbi(s, tcg_rt); + gen_helper_dc_zva(cpu_env, clean_addr); + + if (s->ata) { + /* Extract the tag from the register to match STZGM. */ + tag = tcg_temp_new_i64(); + tcg_gen_shri_i64(tag, tcg_rt, 56); + gen_helper_stzgm_tags(cpu_env, clean_addr, tag); + tcg_temp_free_i64(tag); + } + } + return; default: break; } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> D1.10 specifies that exception handlers begin with tag checks overridden. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-41-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 3 +++ 1 file changed, 3 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 @@ static void arm_cpu_do_interrupt_aarch64(CPUState *cs) break; } } + if (cpu_isar_feature(aa64_mte, cpu)) { + new_mode |= PSTATE_TCO; + } pstate_write(env, PSTATE_DAIF | new_mode); env->aarch64 = 1; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> We need to check the memattr of a page in order to determine whether it is Tagged for MTE. Between Stage1 and Stage2, this becomes simpler if we always collect this data, instead of occasionally being presented with NULL. Use the nonnull attribute to allow the compiler to check that all pointer arguments are non-null. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-42-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/internals.h | 3 ++- target/arm/helper.c | 60 ++++++++++++++++++++--------------------- target/arm/m_helper.c | 11 +++++--- target/arm/tlb_helper.c | 4 ++- 4 files changed, 42 insertions(+), 36 deletions(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ bool get_phys_addr(CPUARMState *env, target_ulong address, MMUAccessType access_type, ARMMMUIdx mmu_idx, hwaddr *phys_ptr, MemTxAttrs *attrs, int *prot, target_ulong *page_size, - ARMMMUFaultInfo *fi, ARMCacheAttrs *cacheattrs); + ARMMMUFaultInfo *fi, ARMCacheAttrs *cacheattrs) + __attribute__((nonnull)); void arm_log_exception(int idx); diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static bool get_phys_addr_lpae(CPUARMState *env, target_ulong address, bool s1_is_el0, hwaddr *phys_ptr, MemTxAttrs *txattrs, int *prot, target_ulong *page_size_ptr, - ARMMMUFaultInfo *fi, ARMCacheAttrs *cacheattrs); + ARMMMUFaultInfo *fi, ARMCacheAttrs *cacheattrs) + __attribute__((nonnull)); #endif static void switch_mode(CPUARMState *env, int mode); @@ -XXX,XX +XXX,XX @@ static bool get_phys_addr_lpae(CPUARMState *env, target_ulong address, arm_tlb_bti_gp(txattrs) = true; } - if (cacheattrs != NULL) { - if (mmu_idx == ARMMMUIdx_Stage2) { - cacheattrs->attrs = convert_stage2_attrs(env, - extract32(attrs, 0, 4)); - } else { - /* Index into MAIR registers for cache attributes */ - uint8_t attrindx = extract32(attrs, 0, 3); - uint64_t mair = env->cp15.mair_el[regime_el(env, mmu_idx)]; - assert(attrindx <= 7); - cacheattrs->attrs = extract64(mair, attrindx * 8, 8); - } - cacheattrs->shareability = extract32(attrs, 6, 2); + if (mmu_idx == ARMMMUIdx_Stage2) { + cacheattrs->attrs = convert_stage2_attrs(env, extract32(attrs, 0, 4)); + } else { + /* Index into MAIR registers for cache attributes */ + uint8_t attrindx = extract32(attrs, 0, 3); + uint64_t mair = env->cp15.mair_el[regime_el(env, mmu_idx)]; + assert(attrindx <= 7); + cacheattrs->attrs = extract64(mair, attrindx * 8, 8); } + cacheattrs->shareability = extract32(attrs, 6, 2); *phys_ptr = descaddr; *page_size_ptr = page_size; @@ -XXX,XX +XXX,XX @@ bool get_phys_addr(CPUARMState *env, target_ulong address, ret = get_phys_addr_lpae(env, ipa, access_type, ARMMMUIdx_Stage2, mmu_idx == ARMMMUIdx_E10_0, phys_ptr, attrs, &s2_prot, - page_size, fi, - cacheattrs != NULL ? &cacheattrs2 : NULL); + page_size, fi, &cacheattrs2); fi->s2addr = ipa; /* Combine the S1 and S2 perms. */ *prot &= s2_prot; - /* Combine the S1 and S2 cache attributes, if needed */ - if (!ret && cacheattrs != NULL) { - if (env->cp15.hcr_el2 & HCR_DC) { - /* - * HCR.DC forces the first stage attributes to - * Normal Non-Shareable, - * Inner Write-Back Read-Allocate Write-Allocate, - * Outer Write-Back Read-Allocate Write-Allocate. - */ - cacheattrs->attrs = 0xff; - cacheattrs->shareability = 0; - } - *cacheattrs = combine_cacheattrs(*cacheattrs, cacheattrs2); + /* If S2 fails, return early. */ + if (ret) { + return ret; } - return ret; + /* Combine the S1 and S2 cache attributes. */ + if (env->cp15.hcr_el2 & HCR_DC) { + /* + * HCR.DC forces the first stage attributes to + * Normal Non-Shareable, + * Inner Write-Back Read-Allocate Write-Allocate, + * Outer Write-Back Read-Allocate Write-Allocate. + */ + cacheattrs->attrs = 0xff; + cacheattrs->shareability = 0; + } + *cacheattrs = combine_cacheattrs(*cacheattrs, cacheattrs2); + return 0; } else { /* * For non-EL2 CPUs a stage1+stage2 translation is just stage 1. @@ -XXX,XX +XXX,XX @@ hwaddr arm_cpu_get_phys_page_attrs_debug(CPUState *cs, vaddr addr, bool ret; ARMMMUFaultInfo fi = {}; ARMMMUIdx mmu_idx = arm_mmu_idx(env); + ARMCacheAttrs cacheattrs = {}; *attrs = (MemTxAttrs) {}; ret = get_phys_addr(env, addr, 0, mmu_idx, &phys_addr, - attrs, &prot, &page_size, &fi, NULL); + attrs, &prot, &page_size, &fi, &cacheattrs); if (ret) { return -1; diff --git a/target/arm/m_helper.c b/target/arm/m_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/m_helper.c +++ b/target/arm/m_helper.c @@ -XXX,XX +XXX,XX @@ static bool v7m_stack_write(ARMCPU *cpu, uint32_t addr, uint32_t value, hwaddr physaddr; int prot; ARMMMUFaultInfo fi = {}; + ARMCacheAttrs cacheattrs = {}; bool secure = mmu_idx & ARM_MMU_IDX_M_S; int exc; bool exc_secure; if (get_phys_addr(env, addr, MMU_DATA_STORE, mmu_idx, &physaddr, - &attrs, &prot, &page_size, &fi, NULL)) { + &attrs, &prot, &page_size, &fi, &cacheattrs)) { /* MPU/SAU lookup failed */ if (fi.type == ARMFault_QEMU_SFault) { if (mode == STACK_LAZYFP) { @@ -XXX,XX +XXX,XX @@ static bool v7m_stack_read(ARMCPU *cpu, uint32_t *dest, uint32_t addr, hwaddr physaddr; int prot; ARMMMUFaultInfo fi = {}; + ARMCacheAttrs cacheattrs = {}; bool secure = mmu_idx & ARM_MMU_IDX_M_S; int exc; bool exc_secure; uint32_t value; if (get_phys_addr(env, addr, MMU_DATA_LOAD, mmu_idx, &physaddr, - &attrs, &prot, &page_size, &fi, NULL)) { + &attrs, &prot, &page_size, &fi, &cacheattrs)) { /* MPU/SAU lookup failed */ if (fi.type == ARMFault_QEMU_SFault) { qemu_log_mask(CPU_LOG_INT, @@ -XXX,XX +XXX,XX @@ static bool v7m_read_half_insn(ARMCPU *cpu, ARMMMUIdx mmu_idx, V8M_SAttributes sattrs = {}; MemTxAttrs attrs = {}; ARMMMUFaultInfo fi = {}; + ARMCacheAttrs cacheattrs = {}; MemTxResult txres; target_ulong page_size; hwaddr physaddr; @@ -XXX,XX +XXX,XX @@ static bool v7m_read_half_insn(ARMCPU *cpu, ARMMMUIdx mmu_idx, "...really SecureFault with SFSR.INVEP\n"); return false; } - if (get_phys_addr(env, addr, MMU_INST_FETCH, mmu_idx, - &physaddr, &attrs, &prot, &page_size, &fi, NULL)) { + if (get_phys_addr(env, addr, MMU_INST_FETCH, mmu_idx, &physaddr, + &attrs, &prot, &page_size, &fi, &cacheattrs)) { /* the MPU lookup failed */ env->v7m.cfsr[env->v7m.secure] |= R_V7M_CFSR_IACCVIOL_MASK; armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_MEM, env->v7m.secure); diff --git a/target/arm/tlb_helper.c b/target/arm/tlb_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tlb_helper.c +++ b/target/arm/tlb_helper.c @@ -XXX,XX +XXX,XX @@ bool arm_cpu_tlb_fill(CPUState *cs, vaddr address, int size, int prot, ret; MemTxAttrs attrs = {}; ARMMMUFaultInfo fi = {}; + ARMCacheAttrs cacheattrs = {}; /* * Walk the page table and (if the mapping exists) add the page @@ -XXX,XX +XXX,XX @@ bool arm_cpu_tlb_fill(CPUState *cs, vaddr address, int size, */ ret = get_phys_addr(&cpu->env, address, access_type, core_to_arm_mmu_idx(&cpu->env, mmu_idx), - &phys_addr, &attrs, &prot, &page_size, &fi, NULL); + &phys_addr, &attrs, &prot, &page_size, + &fi, &cacheattrs); if (likely(!ret)) { /* * Map a single [sub]page. Regions smaller than our declared -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> This "bit" is a particular value of the page's MemAttr. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-43-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 48 ++++++++++++++++++++++++++++++++++++++--- target/arm/tlb_helper.c | 5 +++++ 2 files changed, 50 insertions(+), 3 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static uint8_t combine_cacheattr_nibble(uint8_t s1, uint8_t s2) */ static ARMCacheAttrs combine_cacheattrs(ARMCacheAttrs s1, ARMCacheAttrs s2) { - uint8_t s1lo = extract32(s1.attrs, 0, 4), s2lo = extract32(s2.attrs, 0, 4); - uint8_t s1hi = extract32(s1.attrs, 4, 4), s2hi = extract32(s2.attrs, 4, 4); + uint8_t s1lo, s2lo, s1hi, s2hi; ARMCacheAttrs ret; + bool tagged = false; + + if (s1.attrs == 0xf0) { + tagged = true; + s1.attrs = 0xff; + } + + s1lo = extract32(s1.attrs, 0, 4); + s2lo = extract32(s2.attrs, 0, 4); + s1hi = extract32(s1.attrs, 4, 4); + s2hi = extract32(s2.attrs, 4, 4); /* Combine shareability attributes (table D4-43) */ if (s1.shareability == 2 || s2.shareability == 2) { @@ -XXX,XX +XXX,XX @@ static ARMCacheAttrs combine_cacheattrs(ARMCacheAttrs s1, ARMCacheAttrs s2) } } + /* TODO: CombineS1S2Desc does not consider transient, only WB, RWA. */ + if (tagged && ret.attrs == 0xff) { + ret.attrs = 0xf0; + } + return ret; } @@ -XXX,XX +XXX,XX @@ bool get_phys_addr(CPUARMState *env, target_ulong address, * Normal Non-Shareable, * Inner Write-Back Read-Allocate Write-Allocate, * Outer Write-Back Read-Allocate Write-Allocate. + * Do not overwrite Tagged within attrs. */ - cacheattrs->attrs = 0xff; + if (cacheattrs->attrs != 0xf0) { + cacheattrs->attrs = 0xff; + } cacheattrs->shareability = 0; } *cacheattrs = combine_cacheattrs(*cacheattrs, cacheattrs2); @@ -XXX,XX +XXX,XX @@ bool get_phys_addr(CPUARMState *env, target_ulong address, /* Definitely a real MMU, not an MPU */ if (regime_translation_disabled(env, mmu_idx)) { + uint64_t hcr; + uint8_t memattr; + /* * MMU disabled. S1 addresses within aa64 translation regimes are * still checked for bounds -- see AArch64.TranslateAddressS1Off. @@ -XXX,XX +XXX,XX @@ bool get_phys_addr(CPUARMState *env, target_ulong address, *phys_ptr = address; *prot = PAGE_READ | PAGE_WRITE | PAGE_EXEC; *page_size = TARGET_PAGE_SIZE; + + /* Fill in cacheattr a-la AArch64.TranslateAddressS1Off. */ + hcr = arm_hcr_el2_eff(env); + cacheattrs->shareability = 0; + if (hcr & HCR_DC) { + if (hcr & HCR_DCT) { + memattr = 0xf0; /* Tagged, Normal, WB, RWA */ + } else { + memattr = 0xff; /* Normal, WB, RWA */ + } + } else if (access_type == MMU_INST_FETCH) { + if (regime_sctlr(env, mmu_idx) & SCTLR_I) { + memattr = 0xee; /* Normal, WT, RA, NT */ + } else { + memattr = 0x44; /* Normal, NC, No */ + } + cacheattrs->shareability = 2; /* outer sharable */ + } else { + memattr = 0x00; /* Device, nGnRnE */ + } + cacheattrs->attrs = memattr; return 0; } diff --git a/target/arm/tlb_helper.c b/target/arm/tlb_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tlb_helper.c +++ b/target/arm/tlb_helper.c @@ -XXX,XX +XXX,XX @@ bool arm_cpu_tlb_fill(CPUState *cs, vaddr address, int size, phys_addr &= TARGET_PAGE_MASK; address &= TARGET_PAGE_MASK; } + /* Notice and record tagged memory. */ + if (cpu_isar_feature(aa64_mte, cpu) && cacheattrs.attrs == 0xf0) { + arm_tlb_mte_tagged(&attrs) = true; + } + tlb_set_page_with_attrs(cs, address, phys_addr, attrs, prot, mmu_idx, page_size); return true; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200626033144.790098-44-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 6 ++++++ hw/arm/virt.c | 55 ++++++++++++++++++++++++++++++++++++++++++++++-- target/arm/cpu.c | 52 +++++++++++++++++++++++++++++++++++++++++---- 3 files changed, 107 insertions(+), 6 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 @@ struct ARMCPU { /* MemoryRegion to use for secure physical accesses */ MemoryRegion *secure_memory; + /* MemoryRegion to use for allocation tag accesses */ + MemoryRegion *tag_memory; + MemoryRegion *secure_tag_memory; + /* For v8M, pointer to the IDAU interface provided by board/SoC */ Object *idau; @@ -XXX,XX +XXX,XX @@ typedef enum ARMMMUIdxBit { typedef enum ARMASIdx { ARMASIdx_NS = 0, ARMASIdx_S = 1, + ARMASIdx_TagNS = 2, + ARMASIdx_TagS = 3, } ARMASIdx; /* Return the Exception Level targeted by debug exceptions. */ diff --git a/hw/arm/virt.c b/hw/arm/virt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/virt.c +++ b/hw/arm/virt.c @@ -XXX,XX +XXX,XX @@ static void create_platform_bus(VirtMachineState *vms) sysbus_mmio_get_region(s, 0)); } +static void create_tag_ram(MemoryRegion *tag_sysmem, + hwaddr base, hwaddr size, + const char *name) +{ + MemoryRegion *tagram = g_new(MemoryRegion, 1); + + memory_region_init_ram(tagram, NULL, name, size / 32, &error_fatal); + memory_region_add_subregion(tag_sysmem, base / 32, tagram); +} + static void create_secure_ram(VirtMachineState *vms, - MemoryRegion *secure_sysmem) + MemoryRegion *secure_sysmem, + MemoryRegion *secure_tag_sysmem) { MemoryRegion *secram = g_new(MemoryRegion, 1); char *nodename; @@ -XXX,XX +XXX,XX @@ static void create_secure_ram(VirtMachineState *vms, qemu_fdt_setprop_string(vms->fdt, nodename, "status", "disabled"); qemu_fdt_setprop_string(vms->fdt, nodename, "secure-status", "okay"); + if (secure_tag_sysmem) { + create_tag_ram(secure_tag_sysmem, base, size, "mach-virt.secure-tag"); + } + g_free(nodename); } @@ -XXX,XX +XXX,XX @@ static void machvirt_init(MachineState *machine) const CPUArchIdList *possible_cpus; MemoryRegion *sysmem = get_system_memory(); MemoryRegion *secure_sysmem = NULL; + MemoryRegion *tag_sysmem = NULL; + MemoryRegion *secure_tag_sysmem = NULL; int n, virt_max_cpus; bool firmware_loaded; bool aarch64 = true; @@ -XXX,XX +XXX,XX @@ static void machvirt_init(MachineState *machine) "secure-memory", &error_abort); } + /* + * The cpu adds the property if and only if MemTag is supported. + * If it is, we must allocate the ram to back that up. + */ + if (object_property_find(cpuobj, "tag-memory", NULL)) { + if (!tag_sysmem) { + tag_sysmem = g_new(MemoryRegion, 1); + memory_region_init(tag_sysmem, OBJECT(machine), + "tag-memory", UINT64_MAX / 32); + + if (vms->secure) { + secure_tag_sysmem = g_new(MemoryRegion, 1); + memory_region_init(secure_tag_sysmem, OBJECT(machine), + "secure-tag-memory", UINT64_MAX / 32); + + /* As with ram, secure-tag takes precedence over tag. */ + memory_region_add_subregion_overlap(secure_tag_sysmem, 0, + tag_sysmem, -1); + } + } + + object_property_set_link(cpuobj, OBJECT(tag_sysmem), + "tag-memory", &error_abort); + if (vms->secure) { + object_property_set_link(cpuobj, OBJECT(secure_tag_sysmem), + "secure-tag-memory", &error_abort); + } + } + qdev_realize(DEVICE(cpuobj), NULL, &error_fatal); object_unref(cpuobj); } @@ -XXX,XX +XXX,XX @@ static void machvirt_init(MachineState *machine) create_uart(vms, VIRT_UART, sysmem, serial_hd(0)); if (vms->secure) { - create_secure_ram(vms, secure_sysmem); + create_secure_ram(vms, secure_sysmem, secure_tag_sysmem); create_uart(vms, VIRT_SECURE_UART, secure_sysmem, serial_hd(1)); } + if (tag_sysmem) { + create_tag_ram(tag_sysmem, vms->memmap[VIRT_MEM].base, + machine->ram_size, "mach-virt.tag"); + } + vms->highmem_ecam &= vms->highmem && (!firmware_loaded || aarch64); create_rtc(vms); 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 @@ void arm_cpu_post_init(Object *obj) if (kvm_enabled()) { kvm_arm_add_vcpu_properties(obj); } + +#ifndef CONFIG_USER_ONLY + if (arm_feature(&cpu->env, ARM_FEATURE_AARCH64) && + cpu_isar_feature(aa64_mte, cpu)) { + object_property_add_link(obj, "tag-memory", + TYPE_MEMORY_REGION, + (Object **)&cpu->tag_memory, + qdev_prop_allow_set_link_before_realize, + OBJ_PROP_LINK_STRONG); + + if (arm_feature(&cpu->env, ARM_FEATURE_EL3)) { + object_property_add_link(obj, "secure-tag-memory", + TYPE_MEMORY_REGION, + (Object **)&cpu->secure_tag_memory, + qdev_prop_allow_set_link_before_realize, + OBJ_PROP_LINK_STRONG); + } + } +#endif } static void arm_cpu_finalizefn(Object *obj) @@ -XXX,XX +XXX,XX @@ static void arm_cpu_realizefn(DeviceState *dev, Error **errp) #ifndef CONFIG_USER_ONLY MachineState *ms = MACHINE(qdev_get_machine()); unsigned int smp_cpus = ms->smp.cpus; + bool has_secure = cpu->has_el3 || arm_feature(env, ARM_FEATURE_M_SECURITY); - if (cpu->has_el3 || arm_feature(env, ARM_FEATURE_M_SECURITY)) { - cs->num_ases = 2; + /* + * We must set cs->num_ases to the final value before + * the first call to cpu_address_space_init. + */ + if (cpu->tag_memory != NULL) { + cs->num_ases = 3 + has_secure; + } else { + cs->num_ases = 1 + has_secure; + } + if (has_secure) { if (!cpu->secure_memory) { cpu->secure_memory = cs->memory; } cpu_address_space_init(cs, ARMASIdx_S, "cpu-secure-memory", cpu->secure_memory); - } else { - cs->num_ases = 1; } + + if (cpu->tag_memory != NULL) { + cpu_address_space_init(cs, ARMASIdx_TagNS, "cpu-tag-memory", + cpu->tag_memory); + if (has_secure) { + cpu_address_space_init(cs, ARMASIdx_TagS, "cpu-tag-memory", + cpu->secure_tag_memory); + } + } else if (cpu_isar_feature(aa64_mte, cpu)) { + /* + * Since there is no tag memory, we can't meaningfully support MTE + * to its fullest. To avoid problems later, when we would come to + * use the tag memory, downgrade support to insns only. + */ + cpu->isar.id_aa64pfr1 = + FIELD_DP64(cpu->isar.id_aa64pfr1, ID_AA64PFR1, MTE, 1); + } + cpu_address_space_init(cs, ARMASIdx_NS, "cpu-memory", cs->memory); /* No core_count specified, default to smp_cpus. */ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Look up the physical address for the given virtual address, convert that to a tag physical address, and finally return the host address that backs it. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-45-richard.henderson@linaro.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/mte_helper.c | 131 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 131 insertions(+) diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ #include "cpu.h" #include "internals.h" #include "exec/exec-all.h" +#include "exec/ram_addr.h" #include "exec/cpu_ldst.h" #include "exec/helper-proto.h" @@ -XXX,XX +XXX,XX @@ static uint8_t *allocation_tag_mem(CPUARMState *env, int ptr_mmu_idx, int ptr_size, MMUAccessType tag_access, int tag_size, uintptr_t ra) { +#ifdef CONFIG_USER_ONLY /* Tag storage not implemented. */ return NULL; +#else + uintptr_t index; + CPUIOTLBEntry *iotlbentry; + int in_page, flags; + ram_addr_t ptr_ra; + hwaddr ptr_paddr, tag_paddr, xlat; + MemoryRegion *mr; + ARMASIdx tag_asi; + AddressSpace *tag_as; + void *host; + + /* + * Probe the first byte of the virtual address. This raises an + * exception for inaccessible pages, and resolves the virtual address + * into the softmmu tlb. + * + * When RA == 0, this is for mte_probe1. The page is expected to be + * valid. Indicate to probe_access_flags no-fault, then assert that + * we received a valid page. + */ + flags = probe_access_flags(env, ptr, ptr_access, ptr_mmu_idx, + ra == 0, &host, ra); + assert(!(flags & TLB_INVALID_MASK)); + + /* + * Find the iotlbentry for ptr. This *must* be present in the TLB + * because we just found the mapping. + * TODO: Perhaps there should be a cputlb helper that returns a + * matching tlb entry + iotlb entry. + */ + index = tlb_index(env, ptr_mmu_idx, ptr); +# ifdef CONFIG_DEBUG_TCG + { + CPUTLBEntry *entry = tlb_entry(env, ptr_mmu_idx, ptr); + target_ulong comparator = (ptr_access == MMU_DATA_LOAD + ? entry->addr_read + : tlb_addr_write(entry)); + g_assert(tlb_hit(comparator, ptr)); + } +# endif + iotlbentry = &env_tlb(env)->d[ptr_mmu_idx].iotlb[index]; + + /* If the virtual page MemAttr != Tagged, access unchecked. */ + if (!arm_tlb_mte_tagged(&iotlbentry->attrs)) { + return NULL; + } + + /* + * If not backed by host ram, there is no tag storage: access unchecked. + * This is probably a guest os bug though, so log it. + */ + if (unlikely(flags & TLB_MMIO)) { + qemu_log_mask(LOG_GUEST_ERROR, + "Page @ 0x%" PRIx64 " indicates Tagged Normal memory " + "but is not backed by host ram\n", ptr); + return NULL; + } + + /* + * The Normal memory access can extend to the next page. E.g. a single + * 8-byte access to the last byte of a page will check only the last + * tag on the first page. + * Any page access exception has priority over tag check exception. + */ + in_page = -(ptr | TARGET_PAGE_MASK); + if (unlikely(ptr_size > in_page)) { + void *ignore; + flags |= probe_access_flags(env, ptr + in_page, ptr_access, + ptr_mmu_idx, ra == 0, &ignore, ra); + assert(!(flags & TLB_INVALID_MASK)); + } + + /* Any debug exception has priority over a tag check exception. */ + if (unlikely(flags & TLB_WATCHPOINT)) { + int wp = ptr_access == MMU_DATA_LOAD ? BP_MEM_READ : BP_MEM_WRITE; + assert(ra != 0); + cpu_check_watchpoint(env_cpu(env), ptr, ptr_size, + iotlbentry->attrs, wp, ra); + } + + /* + * Find the physical address within the normal mem space. + * The memory region lookup must succeed because TLB_MMIO was + * not set in the cputlb lookup above. + */ + mr = memory_region_from_host(host, &ptr_ra); + tcg_debug_assert(mr != NULL); + tcg_debug_assert(memory_region_is_ram(mr)); + ptr_paddr = ptr_ra; + do { + ptr_paddr += mr->addr; + mr = mr->container; + } while (mr); + + /* Convert to the physical address in tag space. */ + tag_paddr = ptr_paddr >> (LOG2_TAG_GRANULE + 1); + + /* Look up the address in tag space. */ + tag_asi = iotlbentry->attrs.secure ? ARMASIdx_TagS : ARMASIdx_TagNS; + tag_as = cpu_get_address_space(env_cpu(env), tag_asi); + mr = address_space_translate(tag_as, tag_paddr, &xlat, NULL, + tag_access == MMU_DATA_STORE, + iotlbentry->attrs); + + /* + * Note that @mr will never be NULL. If there is nothing in the address + * space at @tag_paddr, the translation will return the unallocated memory + * region. For our purposes, the result must be ram. + */ + if (unlikely(!memory_region_is_ram(mr))) { + /* ??? Failure is a board configuration error. */ + qemu_log_mask(LOG_UNIMP, + "Tag Memory @ 0x%" HWADDR_PRIx " not found for " + "Normal Memory @ 0x%" HWADDR_PRIx "\n", + tag_paddr, ptr_paddr); + return NULL; + } + + /* + * Ensure the tag memory is dirty on write, for migration. + * Tag memory can never contain code or display memory (vga). + */ + if (tag_access == MMU_DATA_STORE) { + ram_addr_t tag_ra = memory_region_get_ram_addr(mr) + xlat; + cpu_physical_memory_set_dirty_flag(tag_ra, DIRTY_MEMORY_MIGRATION); + } + + return memory_region_get_ram_ptr(mr) + xlat; +#endif } uint64_t HELPER(irg)(CPUARMState *env, uint64_t rn, uint64_t rm) -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> We now implement all of the components of MTE, without actually supporting any tagged memory. All MTE instructions will work, trivially, so we can enable support. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-46-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu64.c | 5 +++++ 1 file changed, 5 insertions(+) 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 = cpu->isar.id_aa64pfr1; t = FIELD_DP64(t, ID_AA64PFR1, BT, 1); + /* + * Begin with full support for MTE; will be downgraded to MTE=1 + * during realize if the board provides no tag memory. + */ + t = FIELD_DP64(t, ID_AA64PFR1, MTE, 2); cpu->isar.id_aa64pfr1 = t; t = cpu->isar.id_aa64mmfr1; -- 2.20.1