Mostly this is patches from me and RTH cleaning up and doing more decodetree conversion for AArch32 Neon. The major new feature is Dongjiu Geng's patchset to report host memory errors to KVM guests; also a new aspeed board from Patrick Williams. thanks -- PMM The following changes since commit 035b448b84f3557206abc44d786c5d3db2638f7d: Merge remote-tracking branch 'remotes/gkurz/tags/9p-next-2020-05-14' into staging (2020-05-14 10:58:30 +0100) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20200514 for you to fetch changes up to e95485f85657be21135c17a9226e297c21e73360: target/arm: Convert NEON VFMA, VFMS 3-reg-same insns to decodetree (2020-05-14 15:03:09 +0100) ---------------------------------------------------------------- target-arm queue: * target/arm: Use correct GDB XML for M-profile cores * target/arm: Code cleanup to use gvec APIs better * aspeed: Add support for the sonorapass-bmc board * target/arm: Support reporting KVM host memory errors to the guest via ACPI notifications * target/arm: Finish conversion of Neon 3-reg-same insns to decodetree ---------------------------------------------------------------- Dongjiu Geng (10): acpi: nvdimm: change NVDIMM_UUID_LE to a common macro hw/arm/virt: Introduce a RAS machine option docs: APEI GHES generation and CPER record description ACPI: Build related register address fields via hardware error fw_cfg blob ACPI: Build Hardware Error Source Table ACPI: Record the Generic Error Status Block address KVM: Move hwpoison page related functions into kvm-all.c ACPI: Record Generic Error Status Block(GESB) table target-arm: kvm64: handle SIGBUS signal from kernel or KVM MAINTAINERS: Add ACPI/HEST/GHES entries Patrick Williams (1): aspeed: Add support for the sonorapass-bmc board Peter Maydell (18): target/arm: Use correct GDB XML for M-profile cores target/arm: Convert Neon 3-reg-same VQRDMLAH/VQRDMLSH to decodetree target/arm: Convert Neon 3-reg-same SHA to decodetree target/arm: Convert Neon 64-bit element 3-reg-same insns target/arm: Convert Neon VHADD 3-reg-same insns target/arm: Convert Neon VABA/VABD 3-reg-same to decodetree target/arm: Convert Neon VRHADD, VHSUB 3-reg-same insns to decodetree target/arm: Convert Neon VQSHL, VRSHL, VQRSHL 3-reg-same insns to decodetree target/arm: Convert Neon VPMAX/VPMIN 3-reg-same insns to decodetree target/arm: Convert Neon VPADD 3-reg-same insns to decodetree target/arm: Convert Neon VQDMULH/VQRDMULH 3-reg-same to decodetree target/arm: Convert Neon VADD, VSUB, VABD 3-reg-same insns to decodetree target/arm: Convert Neon VPMIN/VPMAX/VPADD float 3-reg-same insns to decodetree target/arm: Convert Neon fp VMUL, VMLA, VMLS 3-reg-same insns to decodetree target/arm: Convert Neon 3-reg-same compare insns to decodetree target/arm: Move 'env' argument of recps_f32 and rsqrts_f32 helpers to usual place target/arm: Convert Neon fp VMAX/VMIN/VMAXNM/VMINNM/VRECPS/VRSQRTS to decodetree target/arm: Convert NEON VFMA, VFMS 3-reg-same insns to decodetree Richard Henderson (16): target/arm: Create gen_gvec_[us]sra target/arm: Create gen_gvec_{u,s}{rshr,rsra} target/arm: Create gen_gvec_{sri,sli} target/arm: Remove unnecessary range check for VSHL target/arm: Tidy handle_vec_simd_shri target/arm: Create gen_gvec_{ceq,clt,cle,cgt,cge}0 target/arm: Create gen_gvec_{mla,mls} target/arm: Swap argument order for VSHL during decode target/arm: Create gen_gvec_{cmtst,ushl,sshl} target/arm: Create gen_gvec_{uqadd, sqadd, uqsub, sqsub} target/arm: Remove fp_status from helper_{recpe, rsqrte}_u32 target/arm: Create gen_gvec_{qrdmla,qrdmls} target/arm: Pass pointer to qc to qrdmla/qrdmls target/arm: Clear tail in gvec_fmul_idx_*, gvec_fmla_idx_* target/arm: Vectorize SABD/UABD target/arm: Vectorize SABA/UABA docs/specs/acpi_hest_ghes.rst | 110 ++ docs/specs/index.rst | 1 + configure | 4 +- default-configs/arm-softmmu.mak | 1 + include/hw/acpi/aml-build.h | 1 + include/hw/acpi/generic_event_device.h | 2 + include/hw/acpi/ghes.h | 74 + include/hw/arm/virt.h | 1 + include/qemu/uuid.h | 27 + include/sysemu/kvm.h | 3 +- include/sysemu/kvm_int.h | 12 + target/arm/cpu.h | 4 + target/arm/helper.h | 78 +- target/arm/internals.h | 5 +- target/arm/translate.h | 84 +- target/i386/cpu.h | 2 + target/arm/neon-dp.decode | 119 +- accel/kvm/kvm-all.c | 36 + hw/acpi/aml-build.c | 2 + hw/acpi/generic_event_device.c | 19 + hw/acpi/ghes.c | 448 ++++++ hw/acpi/nvdimm.c | 10 +- hw/arm/aspeed.c | 78 ++ hw/arm/virt-acpi-build.c | 15 + hw/arm/virt.c | 23 + target/arm/cpu_tcg.c | 1 + target/arm/gdbstub.c | 22 +- target/arm/helper.c | 2 +- target/arm/kvm64.c | 77 ++ target/arm/neon_helper.c | 17 - target/arm/tlb_helper.c | 2 +- target/arm/translate-a64.c | 210 +-- target/arm/translate-neon.inc.c | 682 +++++++++- target/arm/translate.c | 2349 +++++++++++++++++--------------- target/arm/vec_helper.c | 240 +++- target/arm/vfp_helper.c | 9 +- target/i386/kvm.c | 36 - MAINTAINERS | 9 + gdb-xml/arm-m-profile.xml | 27 + hw/acpi/Kconfig | 4 + hw/acpi/Makefile.objs | 1 + 41 files changed, 3402 insertions(+), 1445 deletions(-) create mode 100644 docs/specs/acpi_hest_ghes.rst create mode 100644 include/hw/acpi/ghes.h create mode 100644 hw/acpi/ghes.c create mode 100644 gdb-xml/arm-m-profile.xml GDB's remote protocol requires M-profile cores to use the feature name 'org.gnu.gdb.arm.m-profile' instead of the 'org.gnu.gdb.arm.core' feature used for A- and R-profile cores. We weren't doing this, which meant GDB treated our M-profile cores like A-profile ones. This mostly doesn't matter, but for instance means that it doesn't correctly handle backtraces where an M-profile exception frame is involved. Ship a copy of GDB's arm-m-profile.xml and use it on the M-profile cores. The integer registers have the same offsets as the arm-core.xml, but register 25 is the M-profile XPSR rather than the A-profile CPSR, so we need to update arm_cpu_gdb_read_register() and arm_cpu_gdb_write_register() to handle XSPR reads and writes. Fixes: https://bugs.launchpad.net/qemu/+bug/1877136 Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200507134755.13997-1-peter.maydell@linaro.org --- configure | 4 ++-- target/arm/cpu_tcg.c | 1 + target/arm/gdbstub.c | 22 ++++++++++++++++++---- gdb-xml/arm-m-profile.xml | 27 +++++++++++++++++++++++++++ 4 files changed, 48 insertions(+), 6 deletions(-) create mode 100644 gdb-xml/arm-m-profile.xml diff --git a/configure b/configure index XXXXXXX..XXXXXXX 100755 --- a/configure +++ b/configure @@ -XXX,XX +XXX,XX @@ case "$target_name" in TARGET_SYSTBL_ABI=common,oabi bflt="yes" mttcg="yes" - gdb_xml_files="arm-core.xml arm-vfp.xml arm-vfp3.xml arm-neon.xml" + gdb_xml_files="arm-core.xml arm-vfp.xml arm-vfp3.xml arm-neon.xml arm-m-profile.xml" ;; aarch64|aarch64_be) TARGET_ARCH=aarch64 TARGET_BASE_ARCH=arm bflt="yes" mttcg="yes" - gdb_xml_files="aarch64-core.xml aarch64-fpu.xml arm-core.xml arm-vfp.xml arm-vfp3.xml arm-neon.xml" + gdb_xml_files="aarch64-core.xml aarch64-fpu.xml arm-core.xml arm-vfp.xml arm-vfp3.xml arm-neon.xml arm-m-profile.xml" ;; cris) ;; diff --git a/target/arm/cpu_tcg.c b/target/arm/cpu_tcg.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu_tcg.c +++ b/target/arm/cpu_tcg.c @@ -XXX,XX +XXX,XX @@ static void arm_v7m_class_init(ObjectClass *oc, void *data) #endif cc->cpu_exec_interrupt = arm_v7m_cpu_exec_interrupt; + cc->gdb_core_xml_file = "arm-m-profile.xml"; } static const ARMCPUInfo arm_tcg_cpus[] = { diff --git a/target/arm/gdbstub.c b/target/arm/gdbstub.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/gdbstub.c +++ b/target/arm/gdbstub.c @@ -XXX,XX +XXX,XX @@ int arm_cpu_gdb_read_register(CPUState *cs, GByteArray *mem_buf, int n) } return gdb_get_reg32(mem_buf, 0); case 25: - /* CPSR */ - return gdb_get_reg32(mem_buf, cpsr_read(env)); + /* CPSR, or XPSR for M-profile */ + if (arm_feature(env, ARM_FEATURE_M)) { + return gdb_get_reg32(mem_buf, xpsr_read(env)); + } else { + return gdb_get_reg32(mem_buf, cpsr_read(env)); + } } /* Unknown register. */ return 0; @@ -XXX,XX +XXX,XX @@ int arm_cpu_gdb_write_register(CPUState *cs, uint8_t *mem_buf, int n) } return 4; case 25: - /* CPSR */ - cpsr_write(env, tmp, 0xffffffff, CPSRWriteByGDBStub); + /* CPSR, or XPSR for M-profile */ + if (arm_feature(env, ARM_FEATURE_M)) { + /* + * Don't allow writing to XPSR.Exception as it can cause + * a transition into or out of handler mode (it's not + * writeable via the MSR insn so this is a reasonable + * restriction). Other fields are safe to update. + */ + xpsr_write(env, tmp, ~XPSR_EXCP); + } else { + cpsr_write(env, tmp, 0xffffffff, CPSRWriteByGDBStub); + } return 4; } /* Unknown register. */ diff --git a/gdb-xml/arm-m-profile.xml b/gdb-xml/arm-m-profile.xml new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/gdb-xml/arm-m-profile.xml @@ -XXX,XX +XXX,XX @@ +<?xml version="1.0"?> +<!-- Copyright (C) 2010-2020 Free Software Foundation, Inc. + + Copying and distribution of this file, with or without modification, + are permitted in any medium without royalty provided the copyright + notice and this notice are preserved. --> + +<!DOCTYPE feature SYSTEM "gdb-target.dtd"> +<feature name="org.gnu.gdb.arm.m-profile"> + <reg name="r0" bitsize="32"/> + <reg name="r1" bitsize="32"/> + <reg name="r2" bitsize="32"/> + <reg name="r3" bitsize="32"/> + <reg name="r4" bitsize="32"/> + <reg name="r5" bitsize="32"/> + <reg name="r6" bitsize="32"/> + <reg name="r7" bitsize="32"/> + <reg name="r8" bitsize="32"/> + <reg name="r9" bitsize="32"/> + <reg name="r10" bitsize="32"/> + <reg name="r11" bitsize="32"/> + <reg name="r12" bitsize="32"/> + <reg name="sp" bitsize="32" type="data_ptr"/> + <reg name="lr" bitsize="32"/> + <reg name="pc" bitsize="32" type="code_ptr"/> + <reg name="xpsr" bitsize="32" regnum="25"/> +</feature> -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> The functions eliminate duplication of the special cases for this operation. They match up with the GVecGen2iFn typedef. Add out-of-line helpers. We got away with only having inline expanders because the neon vector size is only 16 bytes, and we know that the inline expansion will always succeed. When we reuse this for SVE, tcg-gvec-op may decide to use an out-of-line helper due to longer vector lengths. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-2-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 10 +++ target/arm/translate.h | 7 +- target/arm/translate-a64.c | 15 +--- target/arm/translate.c | 161 ++++++++++++++++++++++--------------- target/arm/vec_helper.c | 25 ++++++ 5 files changed, 139 insertions(+), 79 deletions(-) diff --git a/target/arm/helper.h b/target/arm/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(gvec_pmull_q, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(neon_pmull_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_ssra_b, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_ssra_h, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_ssra_s, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_ssra_d, TCG_CALL_NO_RWG, void, ptr, ptr, i32) + +DEF_HELPER_FLAGS_3(gvec_usra_b, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_usra_h, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_usra_s, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_usra_d, TCG_CALL_NO_RWG, void, ptr, ptr, i32) + #ifdef TARGET_AARCH64 #include "helper-a64.h" #include "helper-sve.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 @@ extern const GVecGen3 mls_op[4]; extern const GVecGen3 cmtst_op[4]; extern const GVecGen3 sshl_op[4]; extern const GVecGen3 ushl_op[4]; -extern const GVecGen2i ssra_op[4]; -extern const GVecGen2i usra_op[4]; extern const GVecGen2i sri_op[4]; extern const GVecGen2i sli_op[4]; extern const GVecGen4 uqadd_op[4]; @@ -XXX,XX +XXX,XX @@ void gen_sshl_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); void gen_ushl_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b); void gen_sshl_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b); +void gen_gvec_ssra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_usra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz); + /* * Forward to the isar_feature_* tests given a DisasContext pointer. */ 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_vec_simd_shri(DisasContext *s, bool is_q, bool is_u, switch (opcode) { case 0x02: /* SSRA / USRA (accumulate) */ - if (is_u) { - /* Shift count same as element size produces zero to add. */ - if (shift == 8 << size) { - goto done; - } - gen_gvec_op2i(s, is_q, rd, rn, shift, &usra_op[size]); - } else { - /* Shift count same as element size produces all sign to add. */ - if (shift == 8 << size) { - shift -= 1; - } - gen_gvec_op2i(s, is_q, rd, rn, shift, &ssra_op[size]); - } + gen_gvec_fn2i(s, is_q, rd, rn, shift, + is_u ? gen_gvec_usra : gen_gvec_ssra, size); return; case 0x08: /* SRI */ /* Shift count same as element size is valid but does nothing. */ 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_ssra_vec(unsigned vece, TCGv_vec d, TCGv_vec a, int64_t sh) tcg_gen_add_vec(vece, d, d, a); } -static const TCGOpcode vecop_list_ssra[] = { - INDEX_op_sari_vec, INDEX_op_add_vec, 0 -}; +void gen_gvec_ssra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_sari_vec, INDEX_op_add_vec, 0 + }; + static const GVecGen2i ops[4] = { + { .fni8 = gen_ssra8_i64, + .fniv = gen_ssra_vec, + .fno = gen_helper_gvec_ssra_b, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_8 }, + { .fni8 = gen_ssra16_i64, + .fniv = gen_ssra_vec, + .fno = gen_helper_gvec_ssra_h, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_16 }, + { .fni4 = gen_ssra32_i32, + .fniv = gen_ssra_vec, + .fno = gen_helper_gvec_ssra_s, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_32 }, + { .fni8 = gen_ssra64_i64, + .fniv = gen_ssra_vec, + .fno = gen_helper_gvec_ssra_b, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_64 }, + }; -const GVecGen2i ssra_op[4] = { - { .fni8 = gen_ssra8_i64, - .fniv = gen_ssra_vec, - .load_dest = true, - .opt_opc = vecop_list_ssra, - .vece = MO_8 }, - { .fni8 = gen_ssra16_i64, - .fniv = gen_ssra_vec, - .load_dest = true, - .opt_opc = vecop_list_ssra, - .vece = MO_16 }, - { .fni4 = gen_ssra32_i32, - .fniv = gen_ssra_vec, - .load_dest = true, - .opt_opc = vecop_list_ssra, - .vece = MO_32 }, - { .fni8 = gen_ssra64_i64, - .fniv = gen_ssra_vec, - .prefer_i64 = TCG_TARGET_REG_BITS == 64, - .opt_opc = vecop_list_ssra, - .load_dest = true, - .vece = MO_64 }, -}; + /* tszimm encoding produces immediates in the range [1..esize]. */ + tcg_debug_assert(shift > 0); + tcg_debug_assert(shift <= (8 << vece)); + + /* + * Shifts larger than the element size are architecturally valid. + * Signed results in all sign bits. + */ + shift = MIN(shift, (8 << vece) - 1); + tcg_gen_gvec_2i(rd_ofs, rm_ofs, opr_sz, max_sz, shift, &ops[vece]); +} static void gen_usra8_i64(TCGv_i64 d, TCGv_i64 a, int64_t shift) { @@ -XXX,XX +XXX,XX @@ static void gen_usra_vec(unsigned vece, TCGv_vec d, TCGv_vec a, int64_t sh) tcg_gen_add_vec(vece, d, d, a); } -static const TCGOpcode vecop_list_usra[] = { - INDEX_op_shri_vec, INDEX_op_add_vec, 0 -}; +void gen_gvec_usra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_shri_vec, INDEX_op_add_vec, 0 + }; + static const GVecGen2i ops[4] = { + { .fni8 = gen_usra8_i64, + .fniv = gen_usra_vec, + .fno = gen_helper_gvec_usra_b, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_8, }, + { .fni8 = gen_usra16_i64, + .fniv = gen_usra_vec, + .fno = gen_helper_gvec_usra_h, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_16, }, + { .fni4 = gen_usra32_i32, + .fniv = gen_usra_vec, + .fno = gen_helper_gvec_usra_s, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_32, }, + { .fni8 = gen_usra64_i64, + .fniv = gen_usra_vec, + .fno = gen_helper_gvec_usra_d, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_64, }, + }; -const GVecGen2i usra_op[4] = { - { .fni8 = gen_usra8_i64, - .fniv = gen_usra_vec, - .load_dest = true, - .opt_opc = vecop_list_usra, - .vece = MO_8, }, - { .fni8 = gen_usra16_i64, - .fniv = gen_usra_vec, - .load_dest = true, - .opt_opc = vecop_list_usra, - .vece = MO_16, }, - { .fni4 = gen_usra32_i32, - .fniv = gen_usra_vec, - .load_dest = true, - .opt_opc = vecop_list_usra, - .vece = MO_32, }, - { .fni8 = gen_usra64_i64, - .fniv = gen_usra_vec, - .prefer_i64 = TCG_TARGET_REG_BITS == 64, - .load_dest = true, - .opt_opc = vecop_list_usra, - .vece = MO_64, }, -}; + /* tszimm encoding produces immediates in the range [1..esize]. */ + tcg_debug_assert(shift > 0); + tcg_debug_assert(shift <= (8 << vece)); + + /* + * Shifts larger than the element size are architecturally valid. + * Unsigned results in all zeros as input to accumulate: nop. + */ + if (shift < (8 << vece)) { + tcg_gen_gvec_2i(rd_ofs, rm_ofs, opr_sz, max_sz, shift, &ops[vece]); + } else { + /* Nop, but we do need to clear the tail. */ + tcg_gen_gvec_mov(vece, rd_ofs, rd_ofs, opr_sz, max_sz); + } +} static void gen_shr8_ins_i64(TCGv_i64 d, TCGv_i64 a, int64_t shift) { @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) case 1: /* VSRA */ /* Right shift comes here negative. */ shift = -shift; - /* Shifts larger than the element size are architecturally - * valid. Unsigned results in all zeros; signed results - * in all sign bits. - */ - if (!u) { - tcg_gen_gvec_2i(rd_ofs, rm_ofs, vec_size, vec_size, - MIN(shift, (8 << size) - 1), - &ssra_op[size]); - } else if (shift >= 8 << size) { - /* rd += 0 */ + if (u) { + gen_gvec_usra(size, rd_ofs, rm_ofs, shift, + vec_size, vec_size); } else { - tcg_gen_gvec_2i(rd_ofs, rm_ofs, vec_size, vec_size, - shift, &usra_op[size]); + gen_gvec_ssra(size, rd_ofs, rm_ofs, shift, + vec_size, vec_size); } return 0; diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vec_helper.c +++ b/target/arm/vec_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(gvec_sqsub_d)(void *vd, void *vq, void *vn, clear_tail(d, oprsz, simd_maxsz(desc)); } + +#define DO_SRA(NAME, TYPE) \ +void HELPER(NAME)(void *vd, void *vn, uint32_t desc) \ +{ \ + intptr_t i, oprsz = simd_oprsz(desc); \ + int shift = simd_data(desc); \ + TYPE *d = vd, *n = vn; \ + for (i = 0; i < oprsz / sizeof(TYPE); i++) { \ + d[i] += n[i] >> shift; \ + } \ + clear_tail(d, oprsz, simd_maxsz(desc)); \ +} + +DO_SRA(gvec_ssra_b, int8_t) +DO_SRA(gvec_ssra_h, int16_t) +DO_SRA(gvec_ssra_s, int32_t) +DO_SRA(gvec_ssra_d, int64_t) + +DO_SRA(gvec_usra_b, uint8_t) +DO_SRA(gvec_usra_h, uint16_t) +DO_SRA(gvec_usra_s, uint32_t) +DO_SRA(gvec_usra_d, uint64_t) + +#undef DO_SRA + /* * Convert float16 to float32, raising no exceptions and * preserving exceptional values, including SNaN. -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Create vectorized versions of handle_shri_with_rndacc for shift+round and shift+round+accumulate. Add out-of-line helpers in preparation for longer vector lengths from SVE. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-3-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 20 ++ target/arm/translate.h | 9 + target/arm/translate-a64.c | 11 +- target/arm/translate.c | 463 +++++++++++++++++++++++++++++++++++-- target/arm/vec_helper.c | 50 ++++ 5 files changed, 527 insertions(+), 26 deletions(-) diff --git a/target/arm/helper.h b/target/arm/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(gvec_usra_h, TCG_CALL_NO_RWG, void, ptr, ptr, i32) DEF_HELPER_FLAGS_3(gvec_usra_s, TCG_CALL_NO_RWG, void, ptr, ptr, i32) DEF_HELPER_FLAGS_3(gvec_usra_d, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_srshr_b, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_srshr_h, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_srshr_s, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_srshr_d, TCG_CALL_NO_RWG, void, ptr, ptr, i32) + +DEF_HELPER_FLAGS_3(gvec_urshr_b, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_urshr_h, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_urshr_s, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_urshr_d, TCG_CALL_NO_RWG, void, ptr, ptr, i32) + +DEF_HELPER_FLAGS_3(gvec_srsra_b, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_srsra_h, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_srsra_s, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_srsra_d, TCG_CALL_NO_RWG, void, ptr, ptr, i32) + +DEF_HELPER_FLAGS_3(gvec_ursra_b, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_ursra_h, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_ursra_s, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_ursra_d, TCG_CALL_NO_RWG, void, ptr, ptr, i32) + #ifdef TARGET_AARCH64 #include "helper-a64.h" #include "helper-sve.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_gvec_ssra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, void gen_gvec_usra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, int64_t shift, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_srshr(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_urshr(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_srsra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_ursra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz); + /* * Forward to the isar_feature_* tests given a DisasContext pointer. */ 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_vec_simd_shri(DisasContext *s, bool is_q, bool is_u, return; case 0x04: /* SRSHR / URSHR (rounding) */ - break; + gen_gvec_fn2i(s, is_q, rd, rn, shift, + is_u ? gen_gvec_urshr : gen_gvec_srshr, size); + return; + case 0x06: /* SRSRA / URSRA (accum + rounding) */ - accumulate = true; - break; + gen_gvec_fn2i(s, is_q, rd, rn, shift, + is_u ? gen_gvec_ursra : gen_gvec_srsra, size); + return; + default: g_assert_not_reached(); } diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ void gen_gvec_usra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, } } +/* + * Shift one less than the requested amount, and the low bit is + * the rounding bit. For the 8 and 16-bit operations, because we + * mask the low bit, we can perform a normal integer shift instead + * of a vector shift. + */ +static void gen_srshr8_i64(TCGv_i64 d, TCGv_i64 a, int64_t sh) +{ + TCGv_i64 t = tcg_temp_new_i64(); + + tcg_gen_shri_i64(t, a, sh - 1); + tcg_gen_andi_i64(t, t, dup_const(MO_8, 1)); + tcg_gen_vec_sar8i_i64(d, a, sh); + tcg_gen_vec_add8_i64(d, d, t); + tcg_temp_free_i64(t); +} + +static void gen_srshr16_i64(TCGv_i64 d, TCGv_i64 a, int64_t sh) +{ + TCGv_i64 t = tcg_temp_new_i64(); + + tcg_gen_shri_i64(t, a, sh - 1); + tcg_gen_andi_i64(t, t, dup_const(MO_16, 1)); + tcg_gen_vec_sar16i_i64(d, a, sh); + tcg_gen_vec_add16_i64(d, d, t); + tcg_temp_free_i64(t); +} + +static void gen_srshr32_i32(TCGv_i32 d, TCGv_i32 a, int32_t sh) +{ + TCGv_i32 t = tcg_temp_new_i32(); + + tcg_gen_extract_i32(t, a, sh - 1, 1); + tcg_gen_sari_i32(d, a, sh); + tcg_gen_add_i32(d, d, t); + tcg_temp_free_i32(t); +} + +static void gen_srshr64_i64(TCGv_i64 d, TCGv_i64 a, int64_t sh) +{ + TCGv_i64 t = tcg_temp_new_i64(); + + tcg_gen_extract_i64(t, a, sh - 1, 1); + tcg_gen_sari_i64(d, a, sh); + tcg_gen_add_i64(d, d, t); + tcg_temp_free_i64(t); +} + +static void gen_srshr_vec(unsigned vece, TCGv_vec d, TCGv_vec a, int64_t sh) +{ + TCGv_vec t = tcg_temp_new_vec_matching(d); + TCGv_vec ones = tcg_temp_new_vec_matching(d); + + tcg_gen_shri_vec(vece, t, a, sh - 1); + tcg_gen_dupi_vec(vece, ones, 1); + tcg_gen_and_vec(vece, t, t, ones); + tcg_gen_sari_vec(vece, d, a, sh); + tcg_gen_add_vec(vece, d, d, t); + + tcg_temp_free_vec(t); + tcg_temp_free_vec(ones); +} + +void gen_gvec_srshr(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_shri_vec, INDEX_op_sari_vec, INDEX_op_add_vec, 0 + }; + static const GVecGen2i ops[4] = { + { .fni8 = gen_srshr8_i64, + .fniv = gen_srshr_vec, + .fno = gen_helper_gvec_srshr_b, + .opt_opc = vecop_list, + .vece = MO_8 }, + { .fni8 = gen_srshr16_i64, + .fniv = gen_srshr_vec, + .fno = gen_helper_gvec_srshr_h, + .opt_opc = vecop_list, + .vece = MO_16 }, + { .fni4 = gen_srshr32_i32, + .fniv = gen_srshr_vec, + .fno = gen_helper_gvec_srshr_s, + .opt_opc = vecop_list, + .vece = MO_32 }, + { .fni8 = gen_srshr64_i64, + .fniv = gen_srshr_vec, + .fno = gen_helper_gvec_srshr_d, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .opt_opc = vecop_list, + .vece = MO_64 }, + }; + + /* tszimm encoding produces immediates in the range [1..esize] */ + tcg_debug_assert(shift > 0); + tcg_debug_assert(shift <= (8 << vece)); + + if (shift == (8 << vece)) { + /* + * Shifts larger than the element size are architecturally valid. + * Signed results in all sign bits. With rounding, this produces + * (-1 + 1) >> 1 == 0, or (0 + 1) >> 1 == 0. + * I.e. always zero. + */ + tcg_gen_gvec_dup_imm(vece, rd_ofs, opr_sz, max_sz, 0); + } else { + tcg_gen_gvec_2i(rd_ofs, rm_ofs, opr_sz, max_sz, shift, &ops[vece]); + } +} + +static void gen_srsra8_i64(TCGv_i64 d, TCGv_i64 a, int64_t sh) +{ + TCGv_i64 t = tcg_temp_new_i64(); + + gen_srshr8_i64(t, a, sh); + tcg_gen_vec_add8_i64(d, d, t); + tcg_temp_free_i64(t); +} + +static void gen_srsra16_i64(TCGv_i64 d, TCGv_i64 a, int64_t sh) +{ + TCGv_i64 t = tcg_temp_new_i64(); + + gen_srshr16_i64(t, a, sh); + tcg_gen_vec_add16_i64(d, d, t); + tcg_temp_free_i64(t); +} + +static void gen_srsra32_i32(TCGv_i32 d, TCGv_i32 a, int32_t sh) +{ + TCGv_i32 t = tcg_temp_new_i32(); + + gen_srshr32_i32(t, a, sh); + tcg_gen_add_i32(d, d, t); + tcg_temp_free_i32(t); +} + +static void gen_srsra64_i64(TCGv_i64 d, TCGv_i64 a, int64_t sh) +{ + TCGv_i64 t = tcg_temp_new_i64(); + + gen_srshr64_i64(t, a, sh); + tcg_gen_add_i64(d, d, t); + tcg_temp_free_i64(t); +} + +static void gen_srsra_vec(unsigned vece, TCGv_vec d, TCGv_vec a, int64_t sh) +{ + TCGv_vec t = tcg_temp_new_vec_matching(d); + + gen_srshr_vec(vece, t, a, sh); + tcg_gen_add_vec(vece, d, d, t); + tcg_temp_free_vec(t); +} + +void gen_gvec_srsra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_shri_vec, INDEX_op_sari_vec, INDEX_op_add_vec, 0 + }; + static const GVecGen2i ops[4] = { + { .fni8 = gen_srsra8_i64, + .fniv = gen_srsra_vec, + .fno = gen_helper_gvec_srsra_b, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_8 }, + { .fni8 = gen_srsra16_i64, + .fniv = gen_srsra_vec, + .fno = gen_helper_gvec_srsra_h, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_16 }, + { .fni4 = gen_srsra32_i32, + .fniv = gen_srsra_vec, + .fno = gen_helper_gvec_srsra_s, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_32 }, + { .fni8 = gen_srsra64_i64, + .fniv = gen_srsra_vec, + .fno = gen_helper_gvec_srsra_d, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_64 }, + }; + + /* tszimm encoding produces immediates in the range [1..esize] */ + tcg_debug_assert(shift > 0); + tcg_debug_assert(shift <= (8 << vece)); + + /* + * Shifts larger than the element size are architecturally valid. + * Signed results in all sign bits. With rounding, this produces + * (-1 + 1) >> 1 == 0, or (0 + 1) >> 1 == 0. + * I.e. always zero. With accumulation, this leaves D unchanged. + */ + if (shift == (8 << vece)) { + /* Nop, but we do need to clear the tail. */ + tcg_gen_gvec_mov(vece, rd_ofs, rd_ofs, opr_sz, max_sz); + } else { + tcg_gen_gvec_2i(rd_ofs, rm_ofs, opr_sz, max_sz, shift, &ops[vece]); + } +} + +static void gen_urshr8_i64(TCGv_i64 d, TCGv_i64 a, int64_t sh) +{ + TCGv_i64 t = tcg_temp_new_i64(); + + tcg_gen_shri_i64(t, a, sh - 1); + tcg_gen_andi_i64(t, t, dup_const(MO_8, 1)); + tcg_gen_vec_shr8i_i64(d, a, sh); + tcg_gen_vec_add8_i64(d, d, t); + tcg_temp_free_i64(t); +} + +static void gen_urshr16_i64(TCGv_i64 d, TCGv_i64 a, int64_t sh) +{ + TCGv_i64 t = tcg_temp_new_i64(); + + tcg_gen_shri_i64(t, a, sh - 1); + tcg_gen_andi_i64(t, t, dup_const(MO_16, 1)); + tcg_gen_vec_shr16i_i64(d, a, sh); + tcg_gen_vec_add16_i64(d, d, t); + tcg_temp_free_i64(t); +} + +static void gen_urshr32_i32(TCGv_i32 d, TCGv_i32 a, int32_t sh) +{ + TCGv_i32 t = tcg_temp_new_i32(); + + tcg_gen_extract_i32(t, a, sh - 1, 1); + tcg_gen_shri_i32(d, a, sh); + tcg_gen_add_i32(d, d, t); + tcg_temp_free_i32(t); +} + +static void gen_urshr64_i64(TCGv_i64 d, TCGv_i64 a, int64_t sh) +{ + TCGv_i64 t = tcg_temp_new_i64(); + + tcg_gen_extract_i64(t, a, sh - 1, 1); + tcg_gen_shri_i64(d, a, sh); + tcg_gen_add_i64(d, d, t); + tcg_temp_free_i64(t); +} + +static void gen_urshr_vec(unsigned vece, TCGv_vec d, TCGv_vec a, int64_t shift) +{ + TCGv_vec t = tcg_temp_new_vec_matching(d); + TCGv_vec ones = tcg_temp_new_vec_matching(d); + + tcg_gen_shri_vec(vece, t, a, shift - 1); + tcg_gen_dupi_vec(vece, ones, 1); + tcg_gen_and_vec(vece, t, t, ones); + tcg_gen_shri_vec(vece, d, a, shift); + tcg_gen_add_vec(vece, d, d, t); + + tcg_temp_free_vec(t); + tcg_temp_free_vec(ones); +} + +void gen_gvec_urshr(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_shri_vec, INDEX_op_add_vec, 0 + }; + static const GVecGen2i ops[4] = { + { .fni8 = gen_urshr8_i64, + .fniv = gen_urshr_vec, + .fno = gen_helper_gvec_urshr_b, + .opt_opc = vecop_list, + .vece = MO_8 }, + { .fni8 = gen_urshr16_i64, + .fniv = gen_urshr_vec, + .fno = gen_helper_gvec_urshr_h, + .opt_opc = vecop_list, + .vece = MO_16 }, + { .fni4 = gen_urshr32_i32, + .fniv = gen_urshr_vec, + .fno = gen_helper_gvec_urshr_s, + .opt_opc = vecop_list, + .vece = MO_32 }, + { .fni8 = gen_urshr64_i64, + .fniv = gen_urshr_vec, + .fno = gen_helper_gvec_urshr_d, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .opt_opc = vecop_list, + .vece = MO_64 }, + }; + + /* tszimm encoding produces immediates in the range [1..esize] */ + tcg_debug_assert(shift > 0); + tcg_debug_assert(shift <= (8 << vece)); + + if (shift == (8 << vece)) { + /* + * Shifts larger than the element size are architecturally valid. + * Unsigned results in zero. With rounding, this produces a + * copy of the most significant bit. + */ + tcg_gen_gvec_shri(vece, rd_ofs, rm_ofs, shift - 1, opr_sz, max_sz); + } else { + tcg_gen_gvec_2i(rd_ofs, rm_ofs, opr_sz, max_sz, shift, &ops[vece]); + } +} + +static void gen_ursra8_i64(TCGv_i64 d, TCGv_i64 a, int64_t sh) +{ + TCGv_i64 t = tcg_temp_new_i64(); + + if (sh == 8) { + tcg_gen_vec_shr8i_i64(t, a, 7); + } else { + gen_urshr8_i64(t, a, sh); + } + tcg_gen_vec_add8_i64(d, d, t); + tcg_temp_free_i64(t); +} + +static void gen_ursra16_i64(TCGv_i64 d, TCGv_i64 a, int64_t sh) +{ + TCGv_i64 t = tcg_temp_new_i64(); + + if (sh == 16) { + tcg_gen_vec_shr16i_i64(t, a, 15); + } else { + gen_urshr16_i64(t, a, sh); + } + tcg_gen_vec_add16_i64(d, d, t); + tcg_temp_free_i64(t); +} + +static void gen_ursra32_i32(TCGv_i32 d, TCGv_i32 a, int32_t sh) +{ + TCGv_i32 t = tcg_temp_new_i32(); + + if (sh == 32) { + tcg_gen_shri_i32(t, a, 31); + } else { + gen_urshr32_i32(t, a, sh); + } + tcg_gen_add_i32(d, d, t); + tcg_temp_free_i32(t); +} + +static void gen_ursra64_i64(TCGv_i64 d, TCGv_i64 a, int64_t sh) +{ + TCGv_i64 t = tcg_temp_new_i64(); + + if (sh == 64) { + tcg_gen_shri_i64(t, a, 63); + } else { + gen_urshr64_i64(t, a, sh); + } + tcg_gen_add_i64(d, d, t); + tcg_temp_free_i64(t); +} + +static void gen_ursra_vec(unsigned vece, TCGv_vec d, TCGv_vec a, int64_t sh) +{ + TCGv_vec t = tcg_temp_new_vec_matching(d); + + if (sh == (8 << vece)) { + tcg_gen_shri_vec(vece, t, a, sh - 1); + } else { + gen_urshr_vec(vece, t, a, sh); + } + tcg_gen_add_vec(vece, d, d, t); + tcg_temp_free_vec(t); +} + +void gen_gvec_ursra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_shri_vec, INDEX_op_add_vec, 0 + }; + static const GVecGen2i ops[4] = { + { .fni8 = gen_ursra8_i64, + .fniv = gen_ursra_vec, + .fno = gen_helper_gvec_ursra_b, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_8 }, + { .fni8 = gen_ursra16_i64, + .fniv = gen_ursra_vec, + .fno = gen_helper_gvec_ursra_h, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_16 }, + { .fni4 = gen_ursra32_i32, + .fniv = gen_ursra_vec, + .fno = gen_helper_gvec_ursra_s, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_32 }, + { .fni8 = gen_ursra64_i64, + .fniv = gen_ursra_vec, + .fno = gen_helper_gvec_ursra_d, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_64 }, + }; + + /* tszimm encoding produces immediates in the range [1..esize] */ + tcg_debug_assert(shift > 0); + tcg_debug_assert(shift <= (8 << vece)); + + tcg_gen_gvec_2i(rd_ofs, rm_ofs, opr_sz, max_sz, shift, &ops[vece]); +} + static void gen_shr8_ins_i64(TCGv_i64 d, TCGv_i64 a, int64_t shift) { uint64_t mask = dup_const(MO_8, 0xff >> shift); @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) } return 0; + case 2: /* VRSHR */ + /* Right shift comes here negative. */ + shift = -shift; + if (u) { + gen_gvec_urshr(size, rd_ofs, rm_ofs, shift, + vec_size, vec_size); + } else { + gen_gvec_srshr(size, rd_ofs, rm_ofs, shift, + vec_size, vec_size); + } + return 0; + + case 3: /* VRSRA */ + /* Right shift comes here negative. */ + shift = -shift; + if (u) { + gen_gvec_ursra(size, rd_ofs, rm_ofs, shift, + vec_size, vec_size); + } else { + gen_gvec_srsra(size, rd_ofs, rm_ofs, shift, + vec_size, vec_size); + } + return 0; + case 4: /* VSRI */ if (!u) { return 1; @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) neon_load_reg64(cpu_V0, rm + pass); tcg_gen_movi_i64(cpu_V1, imm); switch (op) { - case 2: /* VRSHR */ - case 3: /* VRSRA */ - if (u) - gen_helper_neon_rshl_u64(cpu_V0, cpu_V0, cpu_V1); - else - gen_helper_neon_rshl_s64(cpu_V0, cpu_V0, cpu_V1); - break; case 6: /* VQSHLU */ gen_helper_neon_qshlu_s64(cpu_V0, cpu_env, cpu_V0, cpu_V1); @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) default: g_assert_not_reached(); } - if (op == 3) { - /* Accumulate. */ - neon_load_reg64(cpu_V1, rd + pass); - tcg_gen_add_i64(cpu_V0, cpu_V0, cpu_V1); - } neon_store_reg64(cpu_V0, rd + pass); } else { /* size < 3 */ /* Operands in T0 and T1. */ @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) tmp2 = tcg_temp_new_i32(); tcg_gen_movi_i32(tmp2, imm); switch (op) { - case 2: /* VRSHR */ - case 3: /* VRSRA */ - GEN_NEON_INTEGER_OP(rshl); - break; case 6: /* VQSHLU */ switch (size) { case 0: @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) g_assert_not_reached(); } tcg_temp_free_i32(tmp2); - - if (op == 3) { - /* Accumulate. */ - tmp2 = neon_load_reg(rd, pass); - gen_neon_add(size, tmp, tmp2); - tcg_temp_free_i32(tmp2); - } neon_store_reg(rd, pass, tmp); } } /* for pass */ diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vec_helper.c +++ b/target/arm/vec_helper.c @@ -XXX,XX +XXX,XX @@ DO_SRA(gvec_usra_d, uint64_t) #undef DO_SRA +#define DO_RSHR(NAME, TYPE) \ +void HELPER(NAME)(void *vd, void *vn, uint32_t desc) \ +{ \ + intptr_t i, oprsz = simd_oprsz(desc); \ + int shift = simd_data(desc); \ + TYPE *d = vd, *n = vn; \ + for (i = 0; i < oprsz / sizeof(TYPE); i++) { \ + TYPE tmp = n[i] >> (shift - 1); \ + d[i] = (tmp >> 1) + (tmp & 1); \ + } \ + clear_tail(d, oprsz, simd_maxsz(desc)); \ +} + +DO_RSHR(gvec_srshr_b, int8_t) +DO_RSHR(gvec_srshr_h, int16_t) +DO_RSHR(gvec_srshr_s, int32_t) +DO_RSHR(gvec_srshr_d, int64_t) + +DO_RSHR(gvec_urshr_b, uint8_t) +DO_RSHR(gvec_urshr_h, uint16_t) +DO_RSHR(gvec_urshr_s, uint32_t) +DO_RSHR(gvec_urshr_d, uint64_t) + +#undef DO_RSHR + +#define DO_RSRA(NAME, TYPE) \ +void HELPER(NAME)(void *vd, void *vn, uint32_t desc) \ +{ \ + intptr_t i, oprsz = simd_oprsz(desc); \ + int shift = simd_data(desc); \ + TYPE *d = vd, *n = vn; \ + for (i = 0; i < oprsz / sizeof(TYPE); i++) { \ + TYPE tmp = n[i] >> (shift - 1); \ + d[i] += (tmp >> 1) + (tmp & 1); \ + } \ + clear_tail(d, oprsz, simd_maxsz(desc)); \ +} + +DO_RSRA(gvec_srsra_b, int8_t) +DO_RSRA(gvec_srsra_h, int16_t) +DO_RSRA(gvec_srsra_s, int32_t) +DO_RSRA(gvec_srsra_d, int64_t) + +DO_RSRA(gvec_ursra_b, uint8_t) +DO_RSRA(gvec_ursra_h, uint16_t) +DO_RSRA(gvec_ursra_s, uint32_t) +DO_RSRA(gvec_ursra_d, uint64_t) + +#undef DO_RSRA + /* * Convert float16 to float32, raising no exceptions and * preserving exceptional values, including SNaN. -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> The functions eliminate duplication of the special cases for this operation. They match up with the GVecGen2iFn typedef. Add out-of-line helpers. We got away with only having inline expanders because the neon vector size is only 16 bytes, and we know that the inline expansion will always succeed. When we reuse this for SVE, tcg-gvec-op may decide to use an out-of-line helper due to longer vector lengths. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-4-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 10 ++ target/arm/translate.h | 7 +- target/arm/translate-a64.c | 20 +--- target/arm/translate.c | 186 +++++++++++++++++++++---------------- target/arm/vec_helper.c | 38 ++++++++ 5 files changed, 160 insertions(+), 101 deletions(-) diff --git a/target/arm/helper.h b/target/arm/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(gvec_ursra_h, TCG_CALL_NO_RWG, void, ptr, ptr, i32) DEF_HELPER_FLAGS_3(gvec_ursra_s, TCG_CALL_NO_RWG, void, ptr, ptr, i32) DEF_HELPER_FLAGS_3(gvec_ursra_d, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_sri_b, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_sri_h, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_sri_s, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_sri_d, TCG_CALL_NO_RWG, void, ptr, ptr, i32) + +DEF_HELPER_FLAGS_3(gvec_sli_b, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_sli_h, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_sli_s, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_3(gvec_sli_d, TCG_CALL_NO_RWG, void, ptr, ptr, i32) + #ifdef TARGET_AARCH64 #include "helper-a64.h" #include "helper-sve.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 @@ extern const GVecGen3 mls_op[4]; extern const GVecGen3 cmtst_op[4]; extern const GVecGen3 sshl_op[4]; extern const GVecGen3 ushl_op[4]; -extern const GVecGen2i sri_op[4]; -extern const GVecGen2i sli_op[4]; extern const GVecGen4 uqadd_op[4]; extern const GVecGen4 sqadd_op[4]; extern const GVecGen4 uqsub_op[4]; @@ -XXX,XX +XXX,XX @@ void gen_gvec_srsra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, void gen_gvec_ursra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, int64_t shift, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_sri(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_sli(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz); + /* * Forward to the isar_feature_* tests given a DisasContext pointer. */ diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void gen_gvec_op2(DisasContext *s, bool is_q, int rd, is_q ? 16 : 8, vec_full_reg_size(s), gvec_op); } -/* Expand a 2-operand + immediate AdvSIMD vector operation using - * an op descriptor. - */ -static void gen_gvec_op2i(DisasContext *s, bool is_q, int rd, - int rn, int64_t imm, const GVecGen2i *gvec_op) -{ - tcg_gen_gvec_2i(vec_full_reg_offset(s, rd), vec_full_reg_offset(s, rn), - is_q ? 16 : 8, vec_full_reg_size(s), imm, gvec_op); -} - /* Expand a 3-operand AdvSIMD vector operation using an op descriptor. */ static void gen_gvec_op3(DisasContext *s, bool is_q, int rd, int rn, int rm, const GVecGen3 *gvec_op) @@ -XXX,XX +XXX,XX @@ static void handle_vec_simd_shri(DisasContext *s, bool is_q, bool is_u, gen_gvec_fn2i(s, is_q, rd, rn, shift, is_u ? gen_gvec_usra : gen_gvec_ssra, size); return; + case 0x08: /* SRI */ - /* Shift count same as element size is valid but does nothing. */ - if (shift == 8 << size) { - goto done; - } - gen_gvec_op2i(s, is_q, rd, rn, shift, &sri_op[size]); + gen_gvec_fn2i(s, is_q, rd, rn, shift, gen_gvec_sri, size); return; case 0x00: /* SSHR / USHR */ @@ -XXX,XX +XXX,XX @@ static void handle_vec_simd_shri(DisasContext *s, bool is_q, bool is_u, } tcg_temp_free_i64(tcg_round); - done: clear_vec_high(s, is_q, rd); } @@ -XXX,XX +XXX,XX @@ static void handle_vec_simd_shli(DisasContext *s, bool is_q, bool insert, } if (insert) { - gen_gvec_op2i(s, is_q, rd, rn, shift, &sli_op[size]); + gen_gvec_fn2i(s, is_q, rd, rn, shift, gen_gvec_sli, size); } else { gen_gvec_fn2i(s, is_q, rd, rn, shift, tcg_gen_gvec_shli, size); } 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_shr64_ins_i64(TCGv_i64 d, TCGv_i64 a, int64_t shift) static void gen_shr_ins_vec(unsigned vece, TCGv_vec d, TCGv_vec a, int64_t sh) { - if (sh == 0) { - tcg_gen_mov_vec(d, a); - } else { - TCGv_vec t = tcg_temp_new_vec_matching(d); - TCGv_vec m = tcg_temp_new_vec_matching(d); + TCGv_vec t = tcg_temp_new_vec_matching(d); + TCGv_vec m = tcg_temp_new_vec_matching(d); - tcg_gen_dupi_vec(vece, m, MAKE_64BIT_MASK((8 << vece) - sh, sh)); - tcg_gen_shri_vec(vece, t, a, sh); - tcg_gen_and_vec(vece, d, d, m); - tcg_gen_or_vec(vece, d, d, t); + tcg_gen_dupi_vec(vece, m, MAKE_64BIT_MASK((8 << vece) - sh, sh)); + tcg_gen_shri_vec(vece, t, a, sh); + tcg_gen_and_vec(vece, d, d, m); + tcg_gen_or_vec(vece, d, d, t); - tcg_temp_free_vec(t); - tcg_temp_free_vec(m); - } + tcg_temp_free_vec(t); + tcg_temp_free_vec(m); } -static const TCGOpcode vecop_list_sri[] = { INDEX_op_shri_vec, 0 }; +void gen_gvec_sri(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { INDEX_op_shri_vec, 0 }; + const GVecGen2i ops[4] = { + { .fni8 = gen_shr8_ins_i64, + .fniv = gen_shr_ins_vec, + .fno = gen_helper_gvec_sri_b, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_8 }, + { .fni8 = gen_shr16_ins_i64, + .fniv = gen_shr_ins_vec, + .fno = gen_helper_gvec_sri_h, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_16 }, + { .fni4 = gen_shr32_ins_i32, + .fniv = gen_shr_ins_vec, + .fno = gen_helper_gvec_sri_s, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_32 }, + { .fni8 = gen_shr64_ins_i64, + .fniv = gen_shr_ins_vec, + .fno = gen_helper_gvec_sri_d, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_64 }, + }; -const GVecGen2i sri_op[4] = { - { .fni8 = gen_shr8_ins_i64, - .fniv = gen_shr_ins_vec, - .load_dest = true, - .opt_opc = vecop_list_sri, - .vece = MO_8 }, - { .fni8 = gen_shr16_ins_i64, - .fniv = gen_shr_ins_vec, - .load_dest = true, - .opt_opc = vecop_list_sri, - .vece = MO_16 }, - { .fni4 = gen_shr32_ins_i32, - .fniv = gen_shr_ins_vec, - .load_dest = true, - .opt_opc = vecop_list_sri, - .vece = MO_32 }, - { .fni8 = gen_shr64_ins_i64, - .fniv = gen_shr_ins_vec, - .prefer_i64 = TCG_TARGET_REG_BITS == 64, - .load_dest = true, - .opt_opc = vecop_list_sri, - .vece = MO_64 }, -}; + /* tszimm encoding produces immediates in the range [1..esize]. */ + tcg_debug_assert(shift > 0); + tcg_debug_assert(shift <= (8 << vece)); + + /* Shift of esize leaves destination unchanged. */ + if (shift < (8 << vece)) { + tcg_gen_gvec_2i(rd_ofs, rm_ofs, opr_sz, max_sz, shift, &ops[vece]); + } else { + /* Nop, but we do need to clear the tail. */ + tcg_gen_gvec_mov(vece, rd_ofs, rd_ofs, opr_sz, max_sz); + } +} static void gen_shl8_ins_i64(TCGv_i64 d, TCGv_i64 a, int64_t shift) { @@ -XXX,XX +XXX,XX @@ static void gen_shl64_ins_i64(TCGv_i64 d, TCGv_i64 a, int64_t shift) static void gen_shl_ins_vec(unsigned vece, TCGv_vec d, TCGv_vec a, int64_t sh) { - if (sh == 0) { - tcg_gen_mov_vec(d, a); - } else { - TCGv_vec t = tcg_temp_new_vec_matching(d); - TCGv_vec m = tcg_temp_new_vec_matching(d); + TCGv_vec t = tcg_temp_new_vec_matching(d); + TCGv_vec m = tcg_temp_new_vec_matching(d); - tcg_gen_dupi_vec(vece, m, MAKE_64BIT_MASK(0, sh)); - tcg_gen_shli_vec(vece, t, a, sh); - tcg_gen_and_vec(vece, d, d, m); - tcg_gen_or_vec(vece, d, d, t); + tcg_gen_shli_vec(vece, t, a, sh); + tcg_gen_dupi_vec(vece, m, MAKE_64BIT_MASK(0, sh)); + tcg_gen_and_vec(vece, d, d, m); + tcg_gen_or_vec(vece, d, d, t); - tcg_temp_free_vec(t); - tcg_temp_free_vec(m); - } + tcg_temp_free_vec(t); + tcg_temp_free_vec(m); } -static const TCGOpcode vecop_list_sli[] = { INDEX_op_shli_vec, 0 }; +void gen_gvec_sli(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + int64_t shift, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { INDEX_op_shli_vec, 0 }; + const GVecGen2i ops[4] = { + { .fni8 = gen_shl8_ins_i64, + .fniv = gen_shl_ins_vec, + .fno = gen_helper_gvec_sli_b, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_8 }, + { .fni8 = gen_shl16_ins_i64, + .fniv = gen_shl_ins_vec, + .fno = gen_helper_gvec_sli_h, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_16 }, + { .fni4 = gen_shl32_ins_i32, + .fniv = gen_shl_ins_vec, + .fno = gen_helper_gvec_sli_s, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_32 }, + { .fni8 = gen_shl64_ins_i64, + .fniv = gen_shl_ins_vec, + .fno = gen_helper_gvec_sli_d, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_64 }, + }; -const GVecGen2i sli_op[4] = { - { .fni8 = gen_shl8_ins_i64, - .fniv = gen_shl_ins_vec, - .load_dest = true, - .opt_opc = vecop_list_sli, - .vece = MO_8 }, - { .fni8 = gen_shl16_ins_i64, - .fniv = gen_shl_ins_vec, - .load_dest = true, - .opt_opc = vecop_list_sli, - .vece = MO_16 }, - { .fni4 = gen_shl32_ins_i32, - .fniv = gen_shl_ins_vec, - .load_dest = true, - .opt_opc = vecop_list_sli, - .vece = MO_32 }, - { .fni8 = gen_shl64_ins_i64, - .fniv = gen_shl_ins_vec, - .prefer_i64 = TCG_TARGET_REG_BITS == 64, - .load_dest = true, - .opt_opc = vecop_list_sli, - .vece = MO_64 }, -}; + /* tszimm encoding produces immediates in the range [0..esize-1]. */ + tcg_debug_assert(shift >= 0); + tcg_debug_assert(shift < (8 << vece)); + + if (shift == 0) { + tcg_gen_gvec_mov(vece, rd_ofs, rm_ofs, opr_sz, max_sz); + } else { + tcg_gen_gvec_2i(rd_ofs, rm_ofs, opr_sz, max_sz, shift, &ops[vece]); + } +} static void gen_mla8_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b) { @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) } /* Right shift comes here negative. */ shift = -shift; - /* Shift out of range leaves destination unchanged. */ - if (shift < 8 << size) { - tcg_gen_gvec_2i(rd_ofs, rm_ofs, vec_size, vec_size, - shift, &sri_op[size]); - } + gen_gvec_sri(size, rd_ofs, rm_ofs, shift, + vec_size, vec_size); return 0; case 5: /* VSHL, VSLI */ if (u) { /* VSLI */ - /* Shift out of range leaves destination unchanged. */ - if (shift < 8 << size) { - tcg_gen_gvec_2i(rd_ofs, rm_ofs, vec_size, - vec_size, shift, &sli_op[size]); - } + gen_gvec_sli(size, rd_ofs, rm_ofs, shift, + vec_size, vec_size); } else { /* VSHL */ /* Shifts larger than the element size are * architecturally valid and results in zero. diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vec_helper.c +++ b/target/arm/vec_helper.c @@ -XXX,XX +XXX,XX @@ DO_RSRA(gvec_ursra_d, uint64_t) #undef DO_RSRA +#define DO_SRI(NAME, TYPE) \ +void HELPER(NAME)(void *vd, void *vn, uint32_t desc) \ +{ \ + intptr_t i, oprsz = simd_oprsz(desc); \ + int shift = simd_data(desc); \ + TYPE *d = vd, *n = vn; \ + for (i = 0; i < oprsz / sizeof(TYPE); i++) { \ + d[i] = deposit64(d[i], 0, sizeof(TYPE) * 8 - shift, n[i] >> shift); \ + } \ + clear_tail(d, oprsz, simd_maxsz(desc)); \ +} + +DO_SRI(gvec_sri_b, uint8_t) +DO_SRI(gvec_sri_h, uint16_t) +DO_SRI(gvec_sri_s, uint32_t) +DO_SRI(gvec_sri_d, uint64_t) + +#undef DO_SRI + +#define DO_SLI(NAME, TYPE) \ +void HELPER(NAME)(void *vd, void *vn, uint32_t desc) \ +{ \ + intptr_t i, oprsz = simd_oprsz(desc); \ + int shift = simd_data(desc); \ + TYPE *d = vd, *n = vn; \ + for (i = 0; i < oprsz / sizeof(TYPE); i++) { \ + d[i] = deposit64(d[i], shift, sizeof(TYPE) * 8 - shift, n[i]); \ + } \ + clear_tail(d, oprsz, simd_maxsz(desc)); \ +} + +DO_SLI(gvec_sli_b, uint8_t) +DO_SLI(gvec_sli_h, uint16_t) +DO_SLI(gvec_sli_s, uint32_t) +DO_SLI(gvec_sli_d, uint64_t) + +#undef DO_SLI + /* * Convert float16 to float32, raising no exceptions and * preserving exceptional values, including SNaN. -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> In 1dc8425e551, while converting to gvec, I added an extra range check against the shift count. This was unnecessary because the encoding of the shift count produces 0 to the element size - 1. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-5-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.c | 12 ++---------- 1 file changed, 2 insertions(+), 10 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_data_insn(DisasContext *s, uint32_t insn) gen_gvec_sli(size, rd_ofs, rm_ofs, shift, vec_size, vec_size); } else { /* VSHL */ - /* Shifts larger than the element size are - * architecturally valid and results in zero. - */ - if (shift >= 8 << size) { - tcg_gen_gvec_dup_imm(size, rd_ofs, - vec_size, vec_size, 0); - } else { - tcg_gen_gvec_shli(size, rd_ofs, rm_ofs, shift, - vec_size, vec_size); - } + tcg_gen_gvec_shli(size, rd_ofs, rm_ofs, shift, + vec_size, vec_size); } return 0; } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Now that we've converted all cases to gvec, there is quite a bit of dead code at the end of the function. Remove it. Sink the call to gen_gvec_fn2i to the end, loading a function pointer within the switch statement. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-6-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 56 ++++++++++---------------------------- 1 file changed, 14 insertions(+), 42 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_vec_simd_shri(DisasContext *s, bool is_q, bool is_u, int size = 32 - clz32(immh) - 1; int immhb = immh << 3 | immb; int shift = 2 * (8 << size) - immhb; - bool accumulate = false; - int dsize = is_q ? 128 : 64; - int esize = 8 << size; - int elements = dsize/esize; - MemOp memop = size | (is_u ? 0 : MO_SIGN); - TCGv_i64 tcg_rn = new_tmp_a64(s); - TCGv_i64 tcg_rd = new_tmp_a64(s); - TCGv_i64 tcg_round; - uint64_t round_const; - int i; + GVecGen2iFn *gvec_fn; if (extract32(immh, 3, 1) && !is_q) { unallocated_encoding(s); @@ -XXX,XX +XXX,XX @@ static void handle_vec_simd_shri(DisasContext *s, bool is_q, bool is_u, switch (opcode) { case 0x02: /* SSRA / USRA (accumulate) */ - gen_gvec_fn2i(s, is_q, rd, rn, shift, - is_u ? gen_gvec_usra : gen_gvec_ssra, size); - return; + gvec_fn = is_u ? gen_gvec_usra : gen_gvec_ssra; + break; case 0x08: /* SRI */ - gen_gvec_fn2i(s, is_q, rd, rn, shift, gen_gvec_sri, size); - return; + gvec_fn = gen_gvec_sri; + break; case 0x00: /* SSHR / USHR */ if (is_u) { @@ -XXX,XX +XXX,XX @@ static void handle_vec_simd_shri(DisasContext *s, bool is_q, bool is_u, /* Shift count the same size as element size produces zero. */ tcg_gen_gvec_dup_imm(size, vec_full_reg_offset(s, rd), is_q ? 16 : 8, vec_full_reg_size(s), 0); - } else { - gen_gvec_fn2i(s, is_q, rd, rn, shift, tcg_gen_gvec_shri, size); + return; } + gvec_fn = tcg_gen_gvec_shri; } else { /* Shift count the same size as element size produces all sign. */ if (shift == 8 << size) { shift -= 1; } - gen_gvec_fn2i(s, is_q, rd, rn, shift, tcg_gen_gvec_sari, size); + gvec_fn = tcg_gen_gvec_sari; } - return; + break; case 0x04: /* SRSHR / URSHR (rounding) */ - gen_gvec_fn2i(s, is_q, rd, rn, shift, - is_u ? gen_gvec_urshr : gen_gvec_srshr, size); - return; + gvec_fn = is_u ? gen_gvec_urshr : gen_gvec_srshr; + break; case 0x06: /* SRSRA / URSRA (accum + rounding) */ - gen_gvec_fn2i(s, is_q, rd, rn, shift, - is_u ? gen_gvec_ursra : gen_gvec_srsra, size); - return; + gvec_fn = is_u ? gen_gvec_ursra : gen_gvec_srsra; + break; default: g_assert_not_reached(); } - round_const = 1ULL << (shift - 1); - tcg_round = tcg_const_i64(round_const); - - for (i = 0; i < elements; i++) { - read_vec_element(s, tcg_rn, rn, i, memop); - if (accumulate) { - read_vec_element(s, tcg_rd, rd, i, memop); - } - - handle_shri_with_rndacc(tcg_rd, tcg_rn, tcg_round, - accumulate, is_u, size, shift); - - write_vec_element(s, tcg_rd, rd, i, size); - } - tcg_temp_free_i64(tcg_round); - - clear_vec_high(s, is_q, rd); + gen_gvec_fn2i(s, is_q, rd, rn, shift, gvec_fn, size); } /* SHL/SLI - Vector shift left */ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Provide a functional interface for the vector expansion. This fits better with the existing set of helpers that we provide for other operations. Macro-ize the 5 nearly identical comparisons. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-7-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.h | 16 ++- target/arm/translate-a64.c | 22 ++-- target/arm/translate.c | 254 ++++++++----------------------------- 3 files changed, 74 insertions(+), 218 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 gen_swstep_exception(DisasContext *s, int isv, int ex) uint64_t vfp_expand_imm(int size, uint8_t imm8); /* Vector operations shared between ARM and AArch64. */ -extern const GVecGen2 ceq0_op[4]; -extern const GVecGen2 clt0_op[4]; -extern const GVecGen2 cgt0_op[4]; -extern const GVecGen2 cle0_op[4]; -extern const GVecGen2 cge0_op[4]; +void gen_gvec_ceq0(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_clt0(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_cgt0(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_cle0(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_cge0(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, + uint32_t opr_sz, uint32_t max_sz); + extern const GVecGen3 mla_op[4]; extern const GVecGen3 mls_op[4]; extern const GVecGen3 cmtst_op[4]; diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void gen_gvec_fn4(DisasContext *s, bool is_q, int rd, int rn, int rm, is_q ? 16 : 8, vec_full_reg_size(s)); } -/* Expand a 2-operand AdvSIMD vector operation using an op descriptor. */ -static void gen_gvec_op2(DisasContext *s, bool is_q, int rd, - int rn, const GVecGen2 *gvec_op) -{ - tcg_gen_gvec_2(vec_full_reg_offset(s, rd), vec_full_reg_offset(s, rn), - is_q ? 16 : 8, vec_full_reg_size(s), gvec_op); -} - /* Expand a 3-operand AdvSIMD vector operation using an op descriptor. */ static void gen_gvec_op3(DisasContext *s, bool is_q, int rd, int rn, int rm, const GVecGen3 *gvec_op) @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc(DisasContext *s, uint32_t insn) } break; case 0x8: /* CMGT, CMGE */ - gen_gvec_op2(s, is_q, rd, rn, u ? &cge0_op[size] : &cgt0_op[size]); + if (u) { + gen_gvec_fn2(s, is_q, rd, rn, gen_gvec_cge0, size); + } else { + gen_gvec_fn2(s, is_q, rd, rn, gen_gvec_cgt0, size); + } return; case 0x9: /* CMEQ, CMLE */ - gen_gvec_op2(s, is_q, rd, rn, u ? &cle0_op[size] : &ceq0_op[size]); + if (u) { + gen_gvec_fn2(s, is_q, rd, rn, gen_gvec_cle0, size); + } else { + gen_gvec_fn2(s, is_q, rd, rn, gen_gvec_ceq0, size); + } return; case 0xa: /* CMLT */ - gen_gvec_op2(s, is_q, rd, rn, &clt0_op[size]); + gen_gvec_fn2(s, is_q, rd, rn, gen_gvec_clt0, size); return; case 0xb: if (u) { /* ABS, NEG */ 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 do_v81_helper(DisasContext *s, gen_helper_gvec_3_ptr *fn, return 1; } -static void gen_ceq0_i32(TCGv_i32 d, TCGv_i32 a) -{ - tcg_gen_setcondi_i32(TCG_COND_EQ, d, a, 0); - tcg_gen_neg_i32(d, d); -} - -static void gen_ceq0_i64(TCGv_i64 d, TCGv_i64 a) -{ - tcg_gen_setcondi_i64(TCG_COND_EQ, d, a, 0); - tcg_gen_neg_i64(d, d); -} - -static void gen_ceq0_vec(unsigned vece, TCGv_vec d, TCGv_vec a) -{ - TCGv_vec zero = tcg_const_zeros_vec_matching(d); - tcg_gen_cmp_vec(TCG_COND_EQ, vece, d, a, zero); - tcg_temp_free_vec(zero); -} +#define GEN_CMP0(NAME, COND) \ + static void gen_##NAME##0_i32(TCGv_i32 d, TCGv_i32 a) \ + { \ + tcg_gen_setcondi_i32(COND, d, a, 0); \ + tcg_gen_neg_i32(d, d); \ + } \ + static void gen_##NAME##0_i64(TCGv_i64 d, TCGv_i64 a) \ + { \ + tcg_gen_setcondi_i64(COND, d, a, 0); \ + tcg_gen_neg_i64(d, d); \ + } \ + static void gen_##NAME##0_vec(unsigned vece, TCGv_vec d, TCGv_vec a) \ + { \ + TCGv_vec zero = tcg_const_zeros_vec_matching(d); \ + tcg_gen_cmp_vec(COND, vece, d, a, zero); \ + tcg_temp_free_vec(zero); \ + } \ + void gen_gvec_##NAME##0(unsigned vece, uint32_t d, uint32_t m, \ + uint32_t opr_sz, uint32_t max_sz) \ + { \ + const GVecGen2 op[4] = { \ + { .fno = gen_helper_gvec_##NAME##0_b, \ + .fniv = gen_##NAME##0_vec, \ + .opt_opc = vecop_list_cmp, \ + .vece = MO_8 }, \ + { .fno = gen_helper_gvec_##NAME##0_h, \ + .fniv = gen_##NAME##0_vec, \ + .opt_opc = vecop_list_cmp, \ + .vece = MO_16 }, \ + { .fni4 = gen_##NAME##0_i32, \ + .fniv = gen_##NAME##0_vec, \ + .opt_opc = vecop_list_cmp, \ + .vece = MO_32 }, \ + { .fni8 = gen_##NAME##0_i64, \ + .fniv = gen_##NAME##0_vec, \ + .opt_opc = vecop_list_cmp, \ + .prefer_i64 = TCG_TARGET_REG_BITS == 64, \ + .vece = MO_64 }, \ + }; \ + tcg_gen_gvec_2(d, m, opr_sz, max_sz, &op[vece]); \ + } static const TCGOpcode vecop_list_cmp[] = { INDEX_op_cmp_vec, 0 }; -const GVecGen2 ceq0_op[4] = { - { .fno = gen_helper_gvec_ceq0_b, - .fniv = gen_ceq0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_8 }, - { .fno = gen_helper_gvec_ceq0_h, - .fniv = gen_ceq0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_16 }, - { .fni4 = gen_ceq0_i32, - .fniv = gen_ceq0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_32 }, - { .fni8 = gen_ceq0_i64, - .fniv = gen_ceq0_vec, - .opt_opc = vecop_list_cmp, - .prefer_i64 = TCG_TARGET_REG_BITS == 64, - .vece = MO_64 }, -}; +GEN_CMP0(ceq, TCG_COND_EQ) +GEN_CMP0(cle, TCG_COND_LE) +GEN_CMP0(cge, TCG_COND_GE) +GEN_CMP0(clt, TCG_COND_LT) +GEN_CMP0(cgt, TCG_COND_GT) -static void gen_cle0_i32(TCGv_i32 d, TCGv_i32 a) -{ - tcg_gen_setcondi_i32(TCG_COND_LE, d, a, 0); - tcg_gen_neg_i32(d, d); -} - -static void gen_cle0_i64(TCGv_i64 d, TCGv_i64 a) -{ - tcg_gen_setcondi_i64(TCG_COND_LE, d, a, 0); - tcg_gen_neg_i64(d, d); -} - -static void gen_cle0_vec(unsigned vece, TCGv_vec d, TCGv_vec a) -{ - TCGv_vec zero = tcg_const_zeros_vec_matching(d); - tcg_gen_cmp_vec(TCG_COND_LE, vece, d, a, zero); - tcg_temp_free_vec(zero); -} - -const GVecGen2 cle0_op[4] = { - { .fno = gen_helper_gvec_cle0_b, - .fniv = gen_cle0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_8 }, - { .fno = gen_helper_gvec_cle0_h, - .fniv = gen_cle0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_16 }, - { .fni4 = gen_cle0_i32, - .fniv = gen_cle0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_32 }, - { .fni8 = gen_cle0_i64, - .fniv = gen_cle0_vec, - .opt_opc = vecop_list_cmp, - .prefer_i64 = TCG_TARGET_REG_BITS == 64, - .vece = MO_64 }, -}; - -static void gen_cge0_i32(TCGv_i32 d, TCGv_i32 a) -{ - tcg_gen_setcondi_i32(TCG_COND_GE, d, a, 0); - tcg_gen_neg_i32(d, d); -} - -static void gen_cge0_i64(TCGv_i64 d, TCGv_i64 a) -{ - tcg_gen_setcondi_i64(TCG_COND_GE, d, a, 0); - tcg_gen_neg_i64(d, d); -} - -static void gen_cge0_vec(unsigned vece, TCGv_vec d, TCGv_vec a) -{ - TCGv_vec zero = tcg_const_zeros_vec_matching(d); - tcg_gen_cmp_vec(TCG_COND_GE, vece, d, a, zero); - tcg_temp_free_vec(zero); -} - -const GVecGen2 cge0_op[4] = { - { .fno = gen_helper_gvec_cge0_b, - .fniv = gen_cge0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_8 }, - { .fno = gen_helper_gvec_cge0_h, - .fniv = gen_cge0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_16 }, - { .fni4 = gen_cge0_i32, - .fniv = gen_cge0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_32 }, - { .fni8 = gen_cge0_i64, - .fniv = gen_cge0_vec, - .opt_opc = vecop_list_cmp, - .prefer_i64 = TCG_TARGET_REG_BITS == 64, - .vece = MO_64 }, -}; - -static void gen_clt0_i32(TCGv_i32 d, TCGv_i32 a) -{ - tcg_gen_setcondi_i32(TCG_COND_LT, d, a, 0); - tcg_gen_neg_i32(d, d); -} - -static void gen_clt0_i64(TCGv_i64 d, TCGv_i64 a) -{ - tcg_gen_setcondi_i64(TCG_COND_LT, d, a, 0); - tcg_gen_neg_i64(d, d); -} - -static void gen_clt0_vec(unsigned vece, TCGv_vec d, TCGv_vec a) -{ - TCGv_vec zero = tcg_const_zeros_vec_matching(d); - tcg_gen_cmp_vec(TCG_COND_LT, vece, d, a, zero); - tcg_temp_free_vec(zero); -} - -const GVecGen2 clt0_op[4] = { - { .fno = gen_helper_gvec_clt0_b, - .fniv = gen_clt0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_8 }, - { .fno = gen_helper_gvec_clt0_h, - .fniv = gen_clt0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_16 }, - { .fni4 = gen_clt0_i32, - .fniv = gen_clt0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_32 }, - { .fni8 = gen_clt0_i64, - .fniv = gen_clt0_vec, - .opt_opc = vecop_list_cmp, - .prefer_i64 = TCG_TARGET_REG_BITS == 64, - .vece = MO_64 }, -}; - -static void gen_cgt0_i32(TCGv_i32 d, TCGv_i32 a) -{ - tcg_gen_setcondi_i32(TCG_COND_GT, d, a, 0); - tcg_gen_neg_i32(d, d); -} - -static void gen_cgt0_i64(TCGv_i64 d, TCGv_i64 a) -{ - tcg_gen_setcondi_i64(TCG_COND_GT, d, a, 0); - tcg_gen_neg_i64(d, d); -} - -static void gen_cgt0_vec(unsigned vece, TCGv_vec d, TCGv_vec a) -{ - TCGv_vec zero = tcg_const_zeros_vec_matching(d); - tcg_gen_cmp_vec(TCG_COND_GT, vece, d, a, zero); - tcg_temp_free_vec(zero); -} - -const GVecGen2 cgt0_op[4] = { - { .fno = gen_helper_gvec_cgt0_b, - .fniv = gen_cgt0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_8 }, - { .fno = gen_helper_gvec_cgt0_h, - .fniv = gen_cgt0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_16 }, - { .fni4 = gen_cgt0_i32, - .fniv = gen_cgt0_vec, - .opt_opc = vecop_list_cmp, - .vece = MO_32 }, - { .fni8 = gen_cgt0_i64, - .fniv = gen_cgt0_vec, - .opt_opc = vecop_list_cmp, - .prefer_i64 = TCG_TARGET_REG_BITS == 64, - .vece = MO_64 }, -}; +#undef GEN_CMP0 static void gen_ssra8_i64(TCGv_i64 d, TCGv_i64 a, int64_t shift) { @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) break; case NEON_2RM_VCEQ0: - tcg_gen_gvec_2(rd_ofs, rm_ofs, vec_size, - vec_size, &ceq0_op[size]); + gen_gvec_ceq0(size, rd_ofs, rm_ofs, vec_size, vec_size); break; case NEON_2RM_VCGT0: - tcg_gen_gvec_2(rd_ofs, rm_ofs, vec_size, - vec_size, &cgt0_op[size]); + gen_gvec_cgt0(size, rd_ofs, rm_ofs, vec_size, vec_size); break; case NEON_2RM_VCLE0: - tcg_gen_gvec_2(rd_ofs, rm_ofs, vec_size, - vec_size, &cle0_op[size]); + gen_gvec_cle0(size, rd_ofs, rm_ofs, vec_size, vec_size); break; case NEON_2RM_VCGE0: - tcg_gen_gvec_2(rd_ofs, rm_ofs, vec_size, - vec_size, &cge0_op[size]); + gen_gvec_cge0(size, rd_ofs, rm_ofs, vec_size, vec_size); break; case NEON_2RM_VCLT0: - tcg_gen_gvec_2(rd_ofs, rm_ofs, vec_size, - vec_size, &clt0_op[size]); + gen_gvec_clt0(size, rd_ofs, rm_ofs, vec_size, vec_size); break; default: -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Provide a functional interface for the vector expansion. This fits better with the existing set of helpers that we provide for other operations. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-8-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.h | 7 +- target/arm/translate-a64.c | 4 +- target/arm/translate-neon.inc.c | 16 +---- target/arm/translate.c | 117 +++++++++++++++++--------------- 4 files changed, 71 insertions(+), 73 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 @@ void gen_gvec_cle0(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, void gen_gvec_cge0(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); -extern const GVecGen3 mla_op[4]; -extern const GVecGen3 mls_op[4]; +void gen_gvec_mla(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_mls(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); + extern const GVecGen3 cmtst_op[4]; extern const GVecGen3 sshl_op[4]; extern const GVecGen3 ushl_op[4]; diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_3same_int(DisasContext *s, uint32_t insn) return; case 0x12: /* MLA, MLS */ if (u) { - gen_gvec_op3(s, is_q, rd, rn, rm, &mls_op[size]); + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_mls, size); } else { - gen_gvec_op3(s, is_q, rd, rn, rm, &mla_op[size]); + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_mla, size); } return; case 0x11: 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_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) +DO_3SAME_NO_SZ_3(VMLA, gen_gvec_mla) +DO_3SAME_NO_SZ_3(VMLS, gen_gvec_mls) #define DO_3SAME_CMP(INSN, COND) \ static void gen_##INSN##_3s(unsigned vece, uint32_t rd_ofs, \ @@ -XXX,XX +XXX,XX @@ static bool trans_VMUL_p_3s(DisasContext *s, arg_3same *a) 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, \ 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_mls_vec(unsigned vece, TCGv_vec d, TCGv_vec a, TCGv_vec b) /* Note that while NEON does not support VMLA and VMLS as 64-bit ops, * these tables are shared with AArch64 which does support them. */ +void gen_gvec_mla(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_mul_vec, INDEX_op_add_vec, 0 + }; + static const GVecGen3 ops[4] = { + { .fni4 = gen_mla8_i32, + .fniv = gen_mla_vec, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_8 }, + { .fni4 = gen_mla16_i32, + .fniv = gen_mla_vec, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_16 }, + { .fni4 = gen_mla32_i32, + .fniv = gen_mla_vec, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_32 }, + { .fni8 = gen_mla64_i64, + .fniv = gen_mla_vec, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_64 }, + }; + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); +} -static const TCGOpcode vecop_list_mla[] = { - INDEX_op_mul_vec, INDEX_op_add_vec, 0 -}; - -static const TCGOpcode vecop_list_mls[] = { - INDEX_op_mul_vec, INDEX_op_sub_vec, 0 -}; - -const GVecGen3 mla_op[4] = { - { .fni4 = gen_mla8_i32, - .fniv = gen_mla_vec, - .load_dest = true, - .opt_opc = vecop_list_mla, - .vece = MO_8 }, - { .fni4 = gen_mla16_i32, - .fniv = gen_mla_vec, - .load_dest = true, - .opt_opc = vecop_list_mla, - .vece = MO_16 }, - { .fni4 = gen_mla32_i32, - .fniv = gen_mla_vec, - .load_dest = true, - .opt_opc = vecop_list_mla, - .vece = MO_32 }, - { .fni8 = gen_mla64_i64, - .fniv = gen_mla_vec, - .prefer_i64 = TCG_TARGET_REG_BITS == 64, - .load_dest = true, - .opt_opc = vecop_list_mla, - .vece = MO_64 }, -}; - -const GVecGen3 mls_op[4] = { - { .fni4 = gen_mls8_i32, - .fniv = gen_mls_vec, - .load_dest = true, - .opt_opc = vecop_list_mls, - .vece = MO_8 }, - { .fni4 = gen_mls16_i32, - .fniv = gen_mls_vec, - .load_dest = true, - .opt_opc = vecop_list_mls, - .vece = MO_16 }, - { .fni4 = gen_mls32_i32, - .fniv = gen_mls_vec, - .load_dest = true, - .opt_opc = vecop_list_mls, - .vece = MO_32 }, - { .fni8 = gen_mls64_i64, - .fniv = gen_mls_vec, - .prefer_i64 = TCG_TARGET_REG_BITS == 64, - .load_dest = true, - .opt_opc = vecop_list_mls, - .vece = MO_64 }, -}; +void gen_gvec_mls(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_mul_vec, INDEX_op_sub_vec, 0 + }; + static const GVecGen3 ops[4] = { + { .fni4 = gen_mls8_i32, + .fniv = gen_mls_vec, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_8 }, + { .fni4 = gen_mls16_i32, + .fniv = gen_mls_vec, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_16 }, + { .fni4 = gen_mls32_i32, + .fniv = gen_mls_vec, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_32 }, + { .fni8 = gen_mls64_i64, + .fniv = gen_mls_vec, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .load_dest = true, + .opt_opc = vecop_list, + .vece = MO_64 }, + }; + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); +} /* CMTST : test is "if (X & Y != 0)". */ static void gen_cmtst_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b) -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Rather than perform the argument swap during code generation, perform it during decode. This means it doesn't have to be special cased later, and we can share code with aarch64 code generation. Hopefully the decode comment addresses any confusion that might arise in between. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-9-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/neon-dp.decode | 17 +++++++++++++++-- target/arm/translate-neon.inc.c | 3 +-- 2 files changed, 16 insertions(+), 4 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 +# The _rev suffix indicates that Vn and Vm are reversed. This is +# the case for shifts. In the Arm ARM these insns are documented +# with the Vm and Vn fields in their usual places, but in the +# assembly the operands are listed "backwards", ie in the order +# Dd, Dm, Dn where other insns use Dd, Dn, Dm. For QEMU we choose +# to consider Vm and Vn as being in different fields in the insn, +# which allows us to avoid special-casing shifts in the trans_ +# function code. We would otherwise need to manually swap the operands +# over to call Neon helper functions that are shared with AArch64, +# which does not have this odd reversed-operand situation. +@3same_rev .... ... . . . size:2 .... .... .... . q:1 . . .... \ + &3same vn=%vm_dp vm=%vn_dp vd=%vd_dp + +VSHL_S_3s 1111 001 0 0 . .. .... .... 0100 . . . 0 .... @3same_rev +VSHL_U_3s 1111 001 1 0 . .. .... .... 0100 . . . 0 .... @3same_rev VMAX_S_3s 1111 001 0 0 . .. .... .... 0110 . . . 0 .... @3same VMAX_U_3s 1111 001 1 0 . .. .... .... 0110 . . . 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_VMUL_p_3s(DisasContext *s, arg_3same *a) 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, \ + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, \ oprsz, maxsz, &OPARRAY[vece]); \ } \ DO_3SAME(INSN, gen_##INSN##_3s) -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Provide a functional interface for the vector expansion. This fits better with the existing set of helpers that we provide for other operations. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-10-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.h | 10 ++- target/arm/translate-a64.c | 18 ++-- target/arm/translate-neon.inc.c | 23 +---- target/arm/translate.c | 146 +++++++++++++++++--------------- 4 files changed, 95 insertions(+), 102 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 @@ void gen_gvec_mla(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, void gen_gvec_mls(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); -extern const GVecGen3 cmtst_op[4]; -extern const GVecGen3 sshl_op[4]; -extern const GVecGen3 ushl_op[4]; +void gen_gvec_cmtst(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_sshl(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_ushl(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); + extern const GVecGen4 uqadd_op[4]; extern const GVecGen4 sqadd_op[4]; extern const GVecGen4 uqsub_op[4]; diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void gen_gvec_fn4(DisasContext *s, bool is_q, int rd, int rn, int rm, is_q ? 16 : 8, vec_full_reg_size(s)); } -/* Expand a 3-operand AdvSIMD vector operation using an op descriptor. */ -static void gen_gvec_op3(DisasContext *s, bool is_q, int rd, - int rn, int rm, const GVecGen3 *gvec_op) -{ - tcg_gen_gvec_3(vec_full_reg_offset(s, rd), vec_full_reg_offset(s, rn), - vec_full_reg_offset(s, rm), is_q ? 16 : 8, - vec_full_reg_size(s), gvec_op); -} - /* Expand a 3-operand operation using an out-of-line helper. */ static void gen_gvec_op3_ool(DisasContext *s, bool is_q, int rd, int rn, int rm, int data, gen_helper_gvec_3 *fn) @@ -XXX,XX +XXX,XX @@ static void disas_simd_3same_int(DisasContext *s, uint32_t insn) (u ? uqsub_op : sqsub_op) + size); return; case 0x08: /* SSHL, USHL */ - gen_gvec_op3(s, is_q, rd, rn, rm, - u ? &ushl_op[size] : &sshl_op[size]); + if (u) { + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_ushl, size); + } else { + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_sshl, size); + } return; case 0x0c: /* SMAX, UMAX */ if (u) { @@ -XXX,XX +XXX,XX @@ static void disas_simd_3same_int(DisasContext *s, uint32_t insn) return; case 0x11: if (!u) { /* CMTST */ - gen_gvec_op3(s, is_q, rd, rn, rm, &cmtst_op[size]); + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_cmtst, size); return; } /* else CMEQ */ 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(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) +DO_3SAME(VSHL_S, gen_gvec_sshl) +DO_3SAME(VSHL_U, gen_gvec_ushl) /* These insns are all gvec_bitsel but with the inputs in various orders. */ #define DO_3SAME_BITSEL(INSN, O1, O2, O3) \ @@ -XXX,XX +XXX,XX @@ DO_3SAME_NO_SZ_3(VMIN_U, tcg_gen_gvec_umin) DO_3SAME_NO_SZ_3(VMUL, tcg_gen_gvec_mul) DO_3SAME_NO_SZ_3(VMLA, gen_gvec_mla) DO_3SAME_NO_SZ_3(VMLS, gen_gvec_mls) +DO_3SAME_NO_SZ_3(VTST, gen_gvec_cmtst) #define DO_3SAME_CMP(INSN, COND) \ static void gen_##INSN##_3s(unsigned vece, uint32_t rd_ofs, \ @@ -XXX,XX +XXX,XX @@ 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) - #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, \ @@ -XXX,XX +XXX,XX @@ static bool trans_VMUL_p_3s(DisasContext *s, arg_3same *a) } return do_3same(s, a, gen_VMUL_p_3s); } - -#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) \ - { \ - tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_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 void gen_cmtst_vec(unsigned vece, TCGv_vec d, TCGv_vec a, TCGv_vec b) tcg_gen_cmp_vec(TCG_COND_NE, vece, d, d, a); } -static const TCGOpcode vecop_list_cmtst[] = { INDEX_op_cmp_vec, 0 }; - -const GVecGen3 cmtst_op[4] = { - { .fni4 = gen_helper_neon_tst_u8, - .fniv = gen_cmtst_vec, - .opt_opc = vecop_list_cmtst, - .vece = MO_8 }, - { .fni4 = gen_helper_neon_tst_u16, - .fniv = gen_cmtst_vec, - .opt_opc = vecop_list_cmtst, - .vece = MO_16 }, - { .fni4 = gen_cmtst_i32, - .fniv = gen_cmtst_vec, - .opt_opc = vecop_list_cmtst, - .vece = MO_32 }, - { .fni8 = gen_cmtst_i64, - .fniv = gen_cmtst_vec, - .prefer_i64 = TCG_TARGET_REG_BITS == 64, - .opt_opc = vecop_list_cmtst, - .vece = MO_64 }, -}; +void gen_gvec_cmtst(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { INDEX_op_cmp_vec, 0 }; + static const GVecGen3 ops[4] = { + { .fni4 = gen_helper_neon_tst_u8, + .fniv = gen_cmtst_vec, + .opt_opc = vecop_list, + .vece = MO_8 }, + { .fni4 = gen_helper_neon_tst_u16, + .fniv = gen_cmtst_vec, + .opt_opc = vecop_list, + .vece = MO_16 }, + { .fni4 = gen_cmtst_i32, + .fniv = gen_cmtst_vec, + .opt_opc = vecop_list, + .vece = MO_32 }, + { .fni8 = gen_cmtst_i64, + .fniv = gen_cmtst_vec, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .opt_opc = vecop_list, + .vece = MO_64 }, + }; + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); +} void gen_ushl_i32(TCGv_i32 dst, TCGv_i32 src, TCGv_i32 shift) { @@ -XXX,XX +XXX,XX @@ static void gen_ushl_vec(unsigned vece, TCGv_vec dst, tcg_temp_free_vec(rsh); } -static const TCGOpcode ushl_list[] = { - INDEX_op_neg_vec, INDEX_op_shlv_vec, - INDEX_op_shrv_vec, INDEX_op_cmp_vec, 0 -}; - -const GVecGen3 ushl_op[4] = { - { .fniv = gen_ushl_vec, - .fno = gen_helper_gvec_ushl_b, - .opt_opc = ushl_list, - .vece = MO_8 }, - { .fniv = gen_ushl_vec, - .fno = gen_helper_gvec_ushl_h, - .opt_opc = ushl_list, - .vece = MO_16 }, - { .fni4 = gen_ushl_i32, - .fniv = gen_ushl_vec, - .opt_opc = ushl_list, - .vece = MO_32 }, - { .fni8 = gen_ushl_i64, - .fniv = gen_ushl_vec, - .opt_opc = ushl_list, - .vece = MO_64 }, -}; +void gen_gvec_ushl(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_neg_vec, INDEX_op_shlv_vec, + INDEX_op_shrv_vec, INDEX_op_cmp_vec, 0 + }; + static const GVecGen3 ops[4] = { + { .fniv = gen_ushl_vec, + .fno = gen_helper_gvec_ushl_b, + .opt_opc = vecop_list, + .vece = MO_8 }, + { .fniv = gen_ushl_vec, + .fno = gen_helper_gvec_ushl_h, + .opt_opc = vecop_list, + .vece = MO_16 }, + { .fni4 = gen_ushl_i32, + .fniv = gen_ushl_vec, + .opt_opc = vecop_list, + .vece = MO_32 }, + { .fni8 = gen_ushl_i64, + .fniv = gen_ushl_vec, + .opt_opc = vecop_list, + .vece = MO_64 }, + }; + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); +} void gen_sshl_i32(TCGv_i32 dst, TCGv_i32 src, TCGv_i32 shift) { @@ -XXX,XX +XXX,XX @@ static void gen_sshl_vec(unsigned vece, TCGv_vec dst, tcg_temp_free_vec(tmp); } -static const TCGOpcode sshl_list[] = { - INDEX_op_neg_vec, INDEX_op_umin_vec, INDEX_op_shlv_vec, - INDEX_op_sarv_vec, INDEX_op_cmp_vec, INDEX_op_cmpsel_vec, 0 -}; - -const GVecGen3 sshl_op[4] = { - { .fniv = gen_sshl_vec, - .fno = gen_helper_gvec_sshl_b, - .opt_opc = sshl_list, - .vece = MO_8 }, - { .fniv = gen_sshl_vec, - .fno = gen_helper_gvec_sshl_h, - .opt_opc = sshl_list, - .vece = MO_16 }, - { .fni4 = gen_sshl_i32, - .fniv = gen_sshl_vec, - .opt_opc = sshl_list, - .vece = MO_32 }, - { .fni8 = gen_sshl_i64, - .fniv = gen_sshl_vec, - .opt_opc = sshl_list, - .vece = MO_64 }, -}; +void gen_gvec_sshl(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_neg_vec, INDEX_op_umin_vec, INDEX_op_shlv_vec, + INDEX_op_sarv_vec, INDEX_op_cmp_vec, INDEX_op_cmpsel_vec, 0 + }; + static const GVecGen3 ops[4] = { + { .fniv = gen_sshl_vec, + .fno = gen_helper_gvec_sshl_b, + .opt_opc = vecop_list, + .vece = MO_8 }, + { .fniv = gen_sshl_vec, + .fno = gen_helper_gvec_sshl_h, + .opt_opc = vecop_list, + .vece = MO_16 }, + { .fni4 = gen_sshl_i32, + .fniv = gen_sshl_vec, + .opt_opc = vecop_list, + .vece = MO_32 }, + { .fni8 = gen_sshl_i64, + .fniv = gen_sshl_vec, + .opt_opc = vecop_list, + .vece = MO_64 }, + }; + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); +} static void gen_uqadd_vec(unsigned vece, TCGv_vec t, TCGv_vec sat, TCGv_vec a, TCGv_vec b) -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Provide a functional interface for the vector expansion. This fits better with the existing set of helpers that we provide for other operations. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-11-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.h | 13 +- target/arm/translate-a64.c | 22 ++- target/arm/translate-neon.inc.c | 19 +-- target/arm/translate.c | 228 +++++++++++++++++--------------- 4 files changed, 147 insertions(+), 135 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 @@ void gen_gvec_sshl(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, void gen_gvec_ushl(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); -extern const GVecGen4 uqadd_op[4]; -extern const GVecGen4 sqadd_op[4]; -extern const GVecGen4 uqsub_op[4]; -extern const GVecGen4 sqsub_op[4]; void gen_cmtst_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b); void gen_ushl_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); void gen_sshl_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); void gen_ushl_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b); void gen_sshl_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b); +void gen_gvec_uqadd_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_sqadd_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_uqsub_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_sqsub_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); + void gen_gvec_ssra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, int64_t shift, uint32_t opr_sz, uint32_t max_sz); void gen_gvec_usra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_3same_int(DisasContext *s, uint32_t insn) switch (opcode) { case 0x01: /* SQADD, UQADD */ - tcg_gen_gvec_4(vec_full_reg_offset(s, rd), - offsetof(CPUARMState, vfp.qc), - vec_full_reg_offset(s, rn), - vec_full_reg_offset(s, rm), - is_q ? 16 : 8, vec_full_reg_size(s), - (u ? uqadd_op : sqadd_op) + size); + if (u) { + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_uqadd_qc, size); + } else { + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_sqadd_qc, size); + } return; case 0x05: /* SQSUB, UQSUB */ - tcg_gen_gvec_4(vec_full_reg_offset(s, rd), - offsetof(CPUARMState, vfp.qc), - vec_full_reg_offset(s, rn), - vec_full_reg_offset(s, rm), - is_q ? 16 : 8, vec_full_reg_size(s), - (u ? uqsub_op : sqsub_op) + size); + if (u) { + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_uqsub_qc, size); + } else { + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_sqsub_qc, size); + } return; case 0x08: /* SSHL, USHL */ if (u) { 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(VORN, tcg_gen_gvec_orc) DO_3SAME(VEOR, tcg_gen_gvec_xor) DO_3SAME(VSHL_S, gen_gvec_sshl) DO_3SAME(VSHL_U, gen_gvec_ushl) +DO_3SAME(VQADD_S, gen_gvec_sqadd_qc) +DO_3SAME(VQADD_U, gen_gvec_uqadd_qc) +DO_3SAME(VQSUB_S, gen_gvec_sqsub_qc) +DO_3SAME(VQSUB_U, gen_gvec_uqsub_qc) /* These insns are all gvec_bitsel but with the inputs in various orders. */ #define DO_3SAME_BITSEL(INSN, O1, O2, O3) \ @@ -XXX,XX +XXX,XX @@ DO_3SAME_CMP(VCGE_S, TCG_COND_GE) DO_3SAME_CMP(VCGE_U, TCG_COND_GEU) DO_3SAME_CMP(VCEQ, TCG_COND_EQ) -#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) - 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) { 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_uqadd_vec(unsigned vece, TCGv_vec t, TCGv_vec sat, tcg_temp_free_vec(x); } -static const TCGOpcode vecop_list_uqadd[] = { - INDEX_op_usadd_vec, INDEX_op_cmp_vec, INDEX_op_add_vec, 0 -}; - -const GVecGen4 uqadd_op[4] = { - { .fniv = gen_uqadd_vec, - .fno = gen_helper_gvec_uqadd_b, - .write_aofs = true, - .opt_opc = vecop_list_uqadd, - .vece = MO_8 }, - { .fniv = gen_uqadd_vec, - .fno = gen_helper_gvec_uqadd_h, - .write_aofs = true, - .opt_opc = vecop_list_uqadd, - .vece = MO_16 }, - { .fniv = gen_uqadd_vec, - .fno = gen_helper_gvec_uqadd_s, - .write_aofs = true, - .opt_opc = vecop_list_uqadd, - .vece = MO_32 }, - { .fniv = gen_uqadd_vec, - .fno = gen_helper_gvec_uqadd_d, - .write_aofs = true, - .opt_opc = vecop_list_uqadd, - .vece = MO_64 }, -}; +void gen_gvec_uqadd_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_usadd_vec, INDEX_op_cmp_vec, INDEX_op_add_vec, 0 + }; + static const GVecGen4 ops[4] = { + { .fniv = gen_uqadd_vec, + .fno = gen_helper_gvec_uqadd_b, + .write_aofs = true, + .opt_opc = vecop_list, + .vece = MO_8 }, + { .fniv = gen_uqadd_vec, + .fno = gen_helper_gvec_uqadd_h, + .write_aofs = true, + .opt_opc = vecop_list, + .vece = MO_16 }, + { .fniv = gen_uqadd_vec, + .fno = gen_helper_gvec_uqadd_s, + .write_aofs = true, + .opt_opc = vecop_list, + .vece = MO_32 }, + { .fniv = gen_uqadd_vec, + .fno = gen_helper_gvec_uqadd_d, + .write_aofs = true, + .opt_opc = vecop_list, + .vece = MO_64 }, + }; + tcg_gen_gvec_4(rd_ofs, offsetof(CPUARMState, vfp.qc), + rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); +} static void gen_sqadd_vec(unsigned vece, TCGv_vec t, TCGv_vec sat, TCGv_vec a, TCGv_vec b) @@ -XXX,XX +XXX,XX @@ static void gen_sqadd_vec(unsigned vece, TCGv_vec t, TCGv_vec sat, tcg_temp_free_vec(x); } -static const TCGOpcode vecop_list_sqadd[] = { - INDEX_op_ssadd_vec, INDEX_op_cmp_vec, INDEX_op_add_vec, 0 -}; - -const GVecGen4 sqadd_op[4] = { - { .fniv = gen_sqadd_vec, - .fno = gen_helper_gvec_sqadd_b, - .opt_opc = vecop_list_sqadd, - .write_aofs = true, - .vece = MO_8 }, - { .fniv = gen_sqadd_vec, - .fno = gen_helper_gvec_sqadd_h, - .opt_opc = vecop_list_sqadd, - .write_aofs = true, - .vece = MO_16 }, - { .fniv = gen_sqadd_vec, - .fno = gen_helper_gvec_sqadd_s, - .opt_opc = vecop_list_sqadd, - .write_aofs = true, - .vece = MO_32 }, - { .fniv = gen_sqadd_vec, - .fno = gen_helper_gvec_sqadd_d, - .opt_opc = vecop_list_sqadd, - .write_aofs = true, - .vece = MO_64 }, -}; +void gen_gvec_sqadd_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_ssadd_vec, INDEX_op_cmp_vec, INDEX_op_add_vec, 0 + }; + static const GVecGen4 ops[4] = { + { .fniv = gen_sqadd_vec, + .fno = gen_helper_gvec_sqadd_b, + .opt_opc = vecop_list, + .write_aofs = true, + .vece = MO_8 }, + { .fniv = gen_sqadd_vec, + .fno = gen_helper_gvec_sqadd_h, + .opt_opc = vecop_list, + .write_aofs = true, + .vece = MO_16 }, + { .fniv = gen_sqadd_vec, + .fno = gen_helper_gvec_sqadd_s, + .opt_opc = vecop_list, + .write_aofs = true, + .vece = MO_32 }, + { .fniv = gen_sqadd_vec, + .fno = gen_helper_gvec_sqadd_d, + .opt_opc = vecop_list, + .write_aofs = true, + .vece = MO_64 }, + }; + tcg_gen_gvec_4(rd_ofs, offsetof(CPUARMState, vfp.qc), + rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); +} static void gen_uqsub_vec(unsigned vece, TCGv_vec t, TCGv_vec sat, TCGv_vec a, TCGv_vec b) @@ -XXX,XX +XXX,XX @@ static void gen_uqsub_vec(unsigned vece, TCGv_vec t, TCGv_vec sat, tcg_temp_free_vec(x); } -static const TCGOpcode vecop_list_uqsub[] = { - INDEX_op_ussub_vec, INDEX_op_cmp_vec, INDEX_op_sub_vec, 0 -}; - -const GVecGen4 uqsub_op[4] = { - { .fniv = gen_uqsub_vec, - .fno = gen_helper_gvec_uqsub_b, - .opt_opc = vecop_list_uqsub, - .write_aofs = true, - .vece = MO_8 }, - { .fniv = gen_uqsub_vec, - .fno = gen_helper_gvec_uqsub_h, - .opt_opc = vecop_list_uqsub, - .write_aofs = true, - .vece = MO_16 }, - { .fniv = gen_uqsub_vec, - .fno = gen_helper_gvec_uqsub_s, - .opt_opc = vecop_list_uqsub, - .write_aofs = true, - .vece = MO_32 }, - { .fniv = gen_uqsub_vec, - .fno = gen_helper_gvec_uqsub_d, - .opt_opc = vecop_list_uqsub, - .write_aofs = true, - .vece = MO_64 }, -}; +void gen_gvec_uqsub_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_ussub_vec, INDEX_op_cmp_vec, INDEX_op_sub_vec, 0 + }; + static const GVecGen4 ops[4] = { + { .fniv = gen_uqsub_vec, + .fno = gen_helper_gvec_uqsub_b, + .opt_opc = vecop_list, + .write_aofs = true, + .vece = MO_8 }, + { .fniv = gen_uqsub_vec, + .fno = gen_helper_gvec_uqsub_h, + .opt_opc = vecop_list, + .write_aofs = true, + .vece = MO_16 }, + { .fniv = gen_uqsub_vec, + .fno = gen_helper_gvec_uqsub_s, + .opt_opc = vecop_list, + .write_aofs = true, + .vece = MO_32 }, + { .fniv = gen_uqsub_vec, + .fno = gen_helper_gvec_uqsub_d, + .opt_opc = vecop_list, + .write_aofs = true, + .vece = MO_64 }, + }; + tcg_gen_gvec_4(rd_ofs, offsetof(CPUARMState, vfp.qc), + rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); +} static void gen_sqsub_vec(unsigned vece, TCGv_vec t, TCGv_vec sat, TCGv_vec a, TCGv_vec b) @@ -XXX,XX +XXX,XX @@ static void gen_sqsub_vec(unsigned vece, TCGv_vec t, TCGv_vec sat, tcg_temp_free_vec(x); } -static const TCGOpcode vecop_list_sqsub[] = { - INDEX_op_sssub_vec, INDEX_op_cmp_vec, INDEX_op_sub_vec, 0 -}; - -const GVecGen4 sqsub_op[4] = { - { .fniv = gen_sqsub_vec, - .fno = gen_helper_gvec_sqsub_b, - .opt_opc = vecop_list_sqsub, - .write_aofs = true, - .vece = MO_8 }, - { .fniv = gen_sqsub_vec, - .fno = gen_helper_gvec_sqsub_h, - .opt_opc = vecop_list_sqsub, - .write_aofs = true, - .vece = MO_16 }, - { .fniv = gen_sqsub_vec, - .fno = gen_helper_gvec_sqsub_s, - .opt_opc = vecop_list_sqsub, - .write_aofs = true, - .vece = MO_32 }, - { .fniv = gen_sqsub_vec, - .fno = gen_helper_gvec_sqsub_d, - .opt_opc = vecop_list_sqsub, - .write_aofs = true, - .vece = MO_64 }, -}; +void gen_gvec_sqsub_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_sssub_vec, INDEX_op_cmp_vec, INDEX_op_sub_vec, 0 + }; + static const GVecGen4 ops[4] = { + { .fniv = gen_sqsub_vec, + .fno = gen_helper_gvec_sqsub_b, + .opt_opc = vecop_list, + .write_aofs = true, + .vece = MO_8 }, + { .fniv = gen_sqsub_vec, + .fno = gen_helper_gvec_sqsub_h, + .opt_opc = vecop_list, + .write_aofs = true, + .vece = MO_16 }, + { .fniv = gen_sqsub_vec, + .fno = gen_helper_gvec_sqsub_s, + .opt_opc = vecop_list, + .write_aofs = true, + .vece = MO_32 }, + { .fniv = gen_sqsub_vec, + .fno = gen_helper_gvec_sqsub_d, + .opt_opc = vecop_list, + .write_aofs = true, + .vece = MO_64 }, + }; + tcg_gen_gvec_4(rd_ofs, offsetof(CPUARMState, vfp.qc), + rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); +} /* Translate a NEON data processing instruction. Return nonzero if the instruction is invalid. -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> These operations do not touch fp_status. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-12-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 4 ++-- target/arm/translate-a64.c | 5 ++--- target/arm/translate.c | 12 ++---------- target/arm/vfp_helper.c | 5 ++--- 4 files changed, 8 insertions(+), 18 deletions(-) diff --git a/target/arm/helper.h b/target/arm/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_2(recpe_f64, TCG_CALL_NO_RWG, f64, f64, ptr) DEF_HELPER_FLAGS_2(rsqrte_f16, TCG_CALL_NO_RWG, f16, f16, ptr) DEF_HELPER_FLAGS_2(rsqrte_f32, TCG_CALL_NO_RWG, f32, f32, ptr) DEF_HELPER_FLAGS_2(rsqrte_f64, TCG_CALL_NO_RWG, f64, f64, ptr) -DEF_HELPER_2(recpe_u32, i32, i32, ptr) -DEF_HELPER_FLAGS_2(rsqrte_u32, TCG_CALL_NO_RWG, i32, i32, ptr) +DEF_HELPER_FLAGS_1(recpe_u32, TCG_CALL_NO_RWG, i32, i32) +DEF_HELPER_FLAGS_1(rsqrte_u32, TCG_CALL_NO_RWG, i32, i32) DEF_HELPER_FLAGS_4(neon_tbl, TCG_CALL_NO_RWG, i32, i32, i32, ptr, i32) DEF_HELPER_3(shl_cc, i32, env, i32, i32) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_2misc_reciprocal(DisasContext *s, int opcode, switch (opcode) { case 0x3c: /* URECPE */ - gen_helper_recpe_u32(tcg_res, tcg_op, fpst); + gen_helper_recpe_u32(tcg_res, tcg_op); break; case 0x3d: /* FRECPE */ gen_helper_recpe_f32(tcg_res, tcg_op, fpst); @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc(DisasContext *s, uint32_t insn) unallocated_encoding(s); return; } - need_fpstatus = true; break; case 0x1e: /* FRINT32Z */ case 0x1f: /* FRINT64Z */ @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc(DisasContext *s, uint32_t insn) gen_helper_rints_exact(tcg_res, tcg_op, tcg_fpstatus); break; case 0x7c: /* URSQRTE */ - gen_helper_rsqrte_u32(tcg_res, tcg_op, tcg_fpstatus); + gen_helper_rsqrte_u32(tcg_res, tcg_op); break; case 0x1e: /* FRINT32Z */ case 0x5e: /* FRINT32X */ 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) break; } case NEON_2RM_VRECPE: - { - TCGv_ptr fpstatus = get_fpstatus_ptr(1); - gen_helper_recpe_u32(tmp, tmp, fpstatus); - tcg_temp_free_ptr(fpstatus); + gen_helper_recpe_u32(tmp, tmp); break; - } case NEON_2RM_VRSQRTE: - { - TCGv_ptr fpstatus = get_fpstatus_ptr(1); - gen_helper_rsqrte_u32(tmp, tmp, fpstatus); - tcg_temp_free_ptr(fpstatus); + gen_helper_rsqrte_u32(tmp, tmp); break; - } case NEON_2RM_VRECPE_F: { TCGv_ptr fpstatus = get_fpstatus_ptr(1); diff --git a/target/arm/vfp_helper.c b/target/arm/vfp_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vfp_helper.c +++ b/target/arm/vfp_helper.c @@ -XXX,XX +XXX,XX @@ float64 HELPER(rsqrte_f64)(float64 input, void *fpstp) return make_float64(val); } -uint32_t HELPER(recpe_u32)(uint32_t a, void *fpstp) +uint32_t HELPER(recpe_u32)(uint32_t a) { - /* float_status *s = fpstp; */ int input, estimate; if ((a & 0x80000000) == 0) { @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(recpe_u32)(uint32_t a, void *fpstp) return deposit32(0, (32 - 9), 9, estimate); } -uint32_t HELPER(rsqrte_u32)(uint32_t a, void *fpstp) +uint32_t HELPER(rsqrte_u32)(uint32_t a) { int estimate; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Provide a functional interface for the vector expansion. This fits better with the existing set of helpers that we provide for other operations. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-13-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.h | 5 ++++ target/arm/translate-a64.c | 34 ++---------------------- target/arm/translate.c | 54 +++++++++++++++++++------------------- 3 files changed, 34 insertions(+), 59 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 @@ void gen_gvec_sri(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, void gen_gvec_sli(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs, int64_t shift, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_sqrdmlah_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_sqrdmlsh_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); + /* * Forward to the isar_feature_* tests given a DisasContext pointer. */ diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void gen_gvec_op3_ool(DisasContext *s, bool is_q, int rd, is_q ? 16 : 8, vec_full_reg_size(s), data, fn); } -/* Expand a 3-operand + env pointer operation using - * an out-of-line helper. - */ -static void gen_gvec_op3_env(DisasContext *s, bool is_q, int rd, - int rn, int rm, gen_helper_gvec_3_ptr *fn) -{ - tcg_gen_gvec_3_ptr(vec_full_reg_offset(s, rd), - vec_full_reg_offset(s, rn), - vec_full_reg_offset(s, rm), cpu_env, - is_q ? 16 : 8, vec_full_reg_size(s), 0, fn); -} - /* Expand a 3-operand + fpstatus pointer + simd data value operation using * an out-of-line helper. */ @@ -XXX,XX +XXX,XX @@ static void disas_simd_three_reg_same_extra(DisasContext *s, uint32_t insn) switch (opcode) { case 0x0: /* SQRDMLAH (vector) */ - switch (size) { - case 1: - gen_gvec_op3_env(s, is_q, rd, rn, rm, gen_helper_gvec_qrdmlah_s16); - break; - case 2: - gen_gvec_op3_env(s, is_q, rd, rn, rm, gen_helper_gvec_qrdmlah_s32); - break; - default: - g_assert_not_reached(); - } + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_sqrdmlah_qc, size); return; case 0x1: /* SQRDMLSH (vector) */ - switch (size) { - case 1: - gen_gvec_op3_env(s, is_q, rd, rn, rm, gen_helper_gvec_qrdmlsh_s16); - break; - case 2: - gen_gvec_op3_env(s, is_q, rd, rn, rm, gen_helper_gvec_qrdmlsh_s32); - break; - default: - g_assert_not_reached(); - } + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_sqrdmlsh_qc, size); return; case 0x2: /* SDOT / UDOT */ 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 const uint8_t neon_2rm_sizes[] = { [NEON_2RM_VCVT_UF] = 0x4, }; - -/* Expand v8.1 simd helper. */ -static int do_v81_helper(DisasContext *s, gen_helper_gvec_3_ptr *fn, - int q, int rd, int rn, int rm) +void gen_gvec_sqrdmlah_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) { - if (dc_isar_feature(aa32_rdm, s)) { - int opr_sz = (1 + q) * 8; - tcg_gen_gvec_3_ptr(vfp_reg_offset(1, rd), - vfp_reg_offset(1, rn), - vfp_reg_offset(1, rm), cpu_env, - opr_sz, opr_sz, 0, fn); - return 0; - } - return 1; + static gen_helper_gvec_3_ptr * const fns[2] = { + gen_helper_gvec_qrdmlah_s16, gen_helper_gvec_qrdmlah_s32 + }; + tcg_debug_assert(vece >= 1 && vece <= 2); + tcg_gen_gvec_3_ptr(rd_ofs, rn_ofs, rm_ofs, cpu_env, + opr_sz, max_sz, 0, fns[vece - 1]); +} + +void gen_gvec_sqrdmlsh_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) +{ + static gen_helper_gvec_3_ptr * const fns[2] = { + gen_helper_gvec_qrdmlsh_s16, gen_helper_gvec_qrdmlsh_s32 + }; + tcg_debug_assert(vece >= 1 && vece <= 2); + tcg_gen_gvec_3_ptr(rd_ofs, rn_ofs, rm_ofs, cpu_env, + opr_sz, max_sz, 0, fns[vece - 1]); } #define GEN_CMP0(NAME, COND) \ @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) break; /* VPADD */ } /* VQRDMLAH */ - switch (size) { - case 1: - return do_v81_helper(s, gen_helper_gvec_qrdmlah_s16, - q, rd, rn, rm); - case 2: - return do_v81_helper(s, gen_helper_gvec_qrdmlah_s32, - q, rd, rn, rm); + if (dc_isar_feature(aa32_rdm, s) && (size == 1 || size == 2)) { + gen_gvec_sqrdmlah_qc(size, rd_ofs, rn_ofs, rm_ofs, + vec_size, vec_size); + return 0; } return 1; @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) break; } /* VQRDMLSH */ - switch (size) { - case 1: - return do_v81_helper(s, gen_helper_gvec_qrdmlsh_s16, - q, rd, rn, rm); - case 2: - return do_v81_helper(s, gen_helper_gvec_qrdmlsh_s32, - q, rd, rn, rm); + if (dc_isar_feature(aa32_rdm, s) && (size == 1 || size == 2)) { + gen_gvec_sqrdmlsh_qc(size, rd_ofs, rn_ofs, rm_ofs, + vec_size, vec_size); + return 0; } return 1; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Pass a pointer directly to env->vfp.qc[0], rather than env. This will allow SVE2, which does not modify QC, to pass a pointer to dummy storage. Change the return type of inl_qrdml.h_s16 to match the sense of the operation: signed. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-14-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.c | 18 ++++++++--- target/arm/vec_helper.c | 70 +++++++++++++++++++++++------------------ 2 files changed, 54 insertions(+), 34 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 const uint8_t neon_2rm_sizes[] = { [NEON_2RM_VCVT_UF] = 0x4, }; +static void gen_gvec_fn3_qc(uint32_t rd_ofs, uint32_t rn_ofs, uint32_t rm_ofs, + uint32_t opr_sz, uint32_t max_sz, + gen_helper_gvec_3_ptr *fn) +{ + TCGv_ptr qc_ptr = tcg_temp_new_ptr(); + + tcg_gen_addi_ptr(qc_ptr, cpu_env, offsetof(CPUARMState, vfp.qc)); + tcg_gen_gvec_3_ptr(rd_ofs, rn_ofs, rm_ofs, qc_ptr, + opr_sz, max_sz, 0, fn); + tcg_temp_free_ptr(qc_ptr); +} + void gen_gvec_sqrdmlah_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) { @@ -XXX,XX +XXX,XX @@ void gen_gvec_sqrdmlah_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, gen_helper_gvec_qrdmlah_s16, gen_helper_gvec_qrdmlah_s32 }; tcg_debug_assert(vece >= 1 && vece <= 2); - tcg_gen_gvec_3_ptr(rd_ofs, rn_ofs, rm_ofs, cpu_env, - opr_sz, max_sz, 0, fns[vece - 1]); + gen_gvec_fn3_qc(rd_ofs, rn_ofs, rm_ofs, opr_sz, max_sz, fns[vece - 1]); } void gen_gvec_sqrdmlsh_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, @@ -XXX,XX +XXX,XX @@ void gen_gvec_sqrdmlsh_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, gen_helper_gvec_qrdmlsh_s16, gen_helper_gvec_qrdmlsh_s32 }; tcg_debug_assert(vece >= 1 && vece <= 2); - tcg_gen_gvec_3_ptr(rd_ofs, rn_ofs, rm_ofs, cpu_env, - opr_sz, max_sz, 0, fns[vece - 1]); + gen_gvec_fn3_qc(rd_ofs, rn_ofs, rm_ofs, opr_sz, max_sz, fns[vece - 1]); } #define GEN_CMP0(NAME, COND) \ diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vec_helper.c +++ b/target/arm/vec_helper.c @@ -XXX,XX +XXX,XX @@ #define H4(x) (x) #endif -#define SET_QC() env->vfp.qc[0] = 1 - static void clear_tail(void *vd, uintptr_t opr_sz, uintptr_t max_sz) { uint64_t *d = vd + opr_sz; @@ -XXX,XX +XXX,XX @@ static void clear_tail(void *vd, uintptr_t opr_sz, uintptr_t max_sz) } /* Signed saturating rounding doubling multiply-accumulate high half, 16-bit */ -static uint16_t inl_qrdmlah_s16(CPUARMState *env, int16_t src1, - int16_t src2, int16_t src3) +static int16_t inl_qrdmlah_s16(int16_t src1, int16_t src2, + int16_t src3, uint32_t *sat) { /* Simplify: * = ((a3 << 16) + ((e1 * e2) << 1) + (1 << 15)) >> 16 @@ -XXX,XX +XXX,XX @@ static uint16_t inl_qrdmlah_s16(CPUARMState *env, int16_t src1, ret = ((int32_t)src3 << 15) + ret + (1 << 14); ret >>= 15; if (ret != (int16_t)ret) { - SET_QC(); + *sat = 1; ret = (ret < 0 ? -0x8000 : 0x7fff); } return ret; @@ -XXX,XX +XXX,XX @@ static uint16_t inl_qrdmlah_s16(CPUARMState *env, int16_t src1, uint32_t HELPER(neon_qrdmlah_s16)(CPUARMState *env, uint32_t src1, uint32_t src2, uint32_t src3) { - uint16_t e1 = inl_qrdmlah_s16(env, src1, src2, src3); - uint16_t e2 = inl_qrdmlah_s16(env, src1 >> 16, src2 >> 16, src3 >> 16); + uint32_t *sat = &env->vfp.qc[0]; + uint16_t e1 = inl_qrdmlah_s16(src1, src2, src3, sat); + uint16_t e2 = inl_qrdmlah_s16(src1 >> 16, src2 >> 16, src3 >> 16, sat); return deposit32(e1, 16, 16, e2); } void HELPER(gvec_qrdmlah_s16)(void *vd, void *vn, void *vm, - void *ve, uint32_t desc) + void *vq, uint32_t desc) { uintptr_t opr_sz = simd_oprsz(desc); int16_t *d = vd; int16_t *n = vn; int16_t *m = vm; - CPUARMState *env = ve; uintptr_t i; for (i = 0; i < opr_sz / 2; ++i) { - d[i] = inl_qrdmlah_s16(env, n[i], m[i], d[i]); + d[i] = inl_qrdmlah_s16(n[i], m[i], d[i], vq); } clear_tail(d, opr_sz, simd_maxsz(desc)); } /* Signed saturating rounding doubling multiply-subtract high half, 16-bit */ -static uint16_t inl_qrdmlsh_s16(CPUARMState *env, int16_t src1, - int16_t src2, int16_t src3) +static int16_t inl_qrdmlsh_s16(int16_t src1, int16_t src2, + int16_t src3, uint32_t *sat) { /* Similarly, using subtraction: * = ((a3 << 16) - ((e1 * e2) << 1) + (1 << 15)) >> 16 @@ -XXX,XX +XXX,XX @@ static uint16_t inl_qrdmlsh_s16(CPUARMState *env, int16_t src1, ret = ((int32_t)src3 << 15) - ret + (1 << 14); ret >>= 15; if (ret != (int16_t)ret) { - SET_QC(); + *sat = 1; ret = (ret < 0 ? -0x8000 : 0x7fff); } return ret; @@ -XXX,XX +XXX,XX @@ static uint16_t inl_qrdmlsh_s16(CPUARMState *env, int16_t src1, uint32_t HELPER(neon_qrdmlsh_s16)(CPUARMState *env, uint32_t src1, uint32_t src2, uint32_t src3) { - uint16_t e1 = inl_qrdmlsh_s16(env, src1, src2, src3); - uint16_t e2 = inl_qrdmlsh_s16(env, src1 >> 16, src2 >> 16, src3 >> 16); + uint32_t *sat = &env->vfp.qc[0]; + uint16_t e1 = inl_qrdmlsh_s16(src1, src2, src3, sat); + uint16_t e2 = inl_qrdmlsh_s16(src1 >> 16, src2 >> 16, src3 >> 16, sat); return deposit32(e1, 16, 16, e2); } void HELPER(gvec_qrdmlsh_s16)(void *vd, void *vn, void *vm, - void *ve, uint32_t desc) + void *vq, uint32_t desc) { uintptr_t opr_sz = simd_oprsz(desc); int16_t *d = vd; int16_t *n = vn; int16_t *m = vm; - CPUARMState *env = ve; uintptr_t i; for (i = 0; i < opr_sz / 2; ++i) { - d[i] = inl_qrdmlsh_s16(env, n[i], m[i], d[i]); + d[i] = inl_qrdmlsh_s16(n[i], m[i], d[i], vq); } clear_tail(d, opr_sz, simd_maxsz(desc)); } /* Signed saturating rounding doubling multiply-accumulate high half, 32-bit */ -uint32_t HELPER(neon_qrdmlah_s32)(CPUARMState *env, int32_t src1, - int32_t src2, int32_t src3) +static int32_t inl_qrdmlah_s32(int32_t src1, int32_t src2, + int32_t src3, uint32_t *sat) { /* Simplify similarly to int_qrdmlah_s16 above. */ int64_t ret = (int64_t)src1 * src2; ret = ((int64_t)src3 << 31) + ret + (1 << 30); ret >>= 31; if (ret != (int32_t)ret) { - SET_QC(); + *sat = 1; ret = (ret < 0 ? INT32_MIN : INT32_MAX); } return ret; } +uint32_t HELPER(neon_qrdmlah_s32)(CPUARMState *env, int32_t src1, + int32_t src2, int32_t src3) +{ + uint32_t *sat = &env->vfp.qc[0]; + return inl_qrdmlah_s32(src1, src2, src3, sat); +} + void HELPER(gvec_qrdmlah_s32)(void *vd, void *vn, void *vm, - void *ve, uint32_t desc) + void *vq, uint32_t desc) { uintptr_t opr_sz = simd_oprsz(desc); int32_t *d = vd; int32_t *n = vn; int32_t *m = vm; - CPUARMState *env = ve; uintptr_t i; for (i = 0; i < opr_sz / 4; ++i) { - d[i] = helper_neon_qrdmlah_s32(env, n[i], m[i], d[i]); + d[i] = inl_qrdmlah_s32(n[i], m[i], d[i], vq); } clear_tail(d, opr_sz, simd_maxsz(desc)); } /* Signed saturating rounding doubling multiply-subtract high half, 32-bit */ -uint32_t HELPER(neon_qrdmlsh_s32)(CPUARMState *env, int32_t src1, - int32_t src2, int32_t src3) +static int32_t inl_qrdmlsh_s32(int32_t src1, int32_t src2, + int32_t src3, uint32_t *sat) { /* Simplify similarly to int_qrdmlsh_s16 above. */ int64_t ret = (int64_t)src1 * src2; ret = ((int64_t)src3 << 31) - ret + (1 << 30); ret >>= 31; if (ret != (int32_t)ret) { - SET_QC(); + *sat = 1; ret = (ret < 0 ? INT32_MIN : INT32_MAX); } return ret; } +uint32_t HELPER(neon_qrdmlsh_s32)(CPUARMState *env, int32_t src1, + int32_t src2, int32_t src3) +{ + uint32_t *sat = &env->vfp.qc[0]; + return inl_qrdmlsh_s32(src1, src2, src3, sat); +} + void HELPER(gvec_qrdmlsh_s32)(void *vd, void *vn, void *vm, - void *ve, uint32_t desc) + void *vq, uint32_t desc) { uintptr_t opr_sz = simd_oprsz(desc); int32_t *d = vd; int32_t *n = vn; int32_t *m = vm; - CPUARMState *env = ve; uintptr_t i; for (i = 0; i < opr_sz / 4; ++i) { - d[i] = helper_neon_qrdmlsh_s32(env, n[i], m[i], d[i]); + d[i] = inl_qrdmlsh_s32(n[i], m[i], d[i], vq); } clear_tail(d, opr_sz, simd_maxsz(desc)); } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Must clear the tail for AdvSIMD when SVE is enabled. Fixes: ca40a6e6e39 Cc: qemu-stable@nongnu.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-15-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/vec_helper.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vec_helper.c +++ b/target/arm/vec_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(NAME)(void *vd, void *vn, void *vm, void *stat, uint32_t desc) \ d[i + j] = TYPE##_mul(n[i + j], mm, stat); \ } \ } \ + clear_tail(d, oprsz, simd_maxsz(desc)); \ } DO_MUL_IDX(gvec_fmul_idx_h, float16, H2) @@ -XXX,XX +XXX,XX @@ void HELPER(NAME)(void *vd, void *vn, void *vm, void *va, \ mm, a[i + j], 0, stat); \ } \ } \ + clear_tail(d, oprsz, simd_maxsz(desc)); \ } DO_FMLA_IDX(gvec_fmla_idx_h, float16, H2) -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Include 64-bit element size in preparation for SVE2. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-16-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 10 +++ target/arm/translate.h | 5 ++ target/arm/translate-a64.c | 8 ++- target/arm/translate.c | 133 ++++++++++++++++++++++++++++++++++++- target/arm/vec_helper.c | 24 +++++++ 5 files changed, 176 insertions(+), 4 deletions(-) diff --git a/target/arm/helper.h b/target/arm/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(gvec_sli_h, TCG_CALL_NO_RWG, void, ptr, ptr, i32) DEF_HELPER_FLAGS_3(gvec_sli_s, TCG_CALL_NO_RWG, void, ptr, ptr, i32) DEF_HELPER_FLAGS_3(gvec_sli_d, TCG_CALL_NO_RWG, void, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_sabd_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_sabd_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_sabd_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_sabd_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(gvec_uabd_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_uabd_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_uabd_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_uabd_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + #ifdef TARGET_AARCH64 #include "helper-a64.h" #include "helper-sve.h" diff --git a/target/arm/translate.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_gvec_sqrdmlah_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, void gen_gvec_sqrdmlsh_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_sabd(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_uabd(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); + /* * Forward to the isar_feature_* tests given a DisasContext pointer. */ diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_3same_int(DisasContext *s, uint32_t insn) gen_gvec_fn3(s, is_q, rd, rn, rm, tcg_gen_gvec_smin, size); } return; + case 0xe: /* SABD, UABD */ + if (u) { + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_uabd, size); + } else { + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_sabd, size); + } + return; case 0x10: /* ADD, SUB */ if (u) { gen_gvec_fn3(s, is_q, rd, rn, rm, tcg_gen_gvec_sub, size); @@ -XXX,XX +XXX,XX @@ static void disas_simd_3same_int(DisasContext *s, uint32_t insn) genenvfn = fns[size][u]; break; } - case 0xe: /* SABD, UABD */ case 0xf: /* SABA, UABA */ { static NeonGenTwoOpFn * const fns[3][2] = { diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ void gen_gvec_sqsub_qc(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); } +static void gen_sabd_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b) +{ + TCGv_i32 t = tcg_temp_new_i32(); + + tcg_gen_sub_i32(t, a, b); + tcg_gen_sub_i32(d, b, a); + tcg_gen_movcond_i32(TCG_COND_LT, d, a, b, d, t); + tcg_temp_free_i32(t); +} + +static void gen_sabd_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b) +{ + TCGv_i64 t = tcg_temp_new_i64(); + + tcg_gen_sub_i64(t, a, b); + tcg_gen_sub_i64(d, b, a); + tcg_gen_movcond_i64(TCG_COND_LT, d, a, b, d, t); + tcg_temp_free_i64(t); +} + +static void gen_sabd_vec(unsigned vece, TCGv_vec d, TCGv_vec a, TCGv_vec b) +{ + TCGv_vec t = tcg_temp_new_vec_matching(d); + + tcg_gen_smin_vec(vece, t, a, b); + tcg_gen_smax_vec(vece, d, a, b); + tcg_gen_sub_vec(vece, d, d, t); + tcg_temp_free_vec(t); +} + +void gen_gvec_sabd(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_sub_vec, INDEX_op_smin_vec, INDEX_op_smax_vec, 0 + }; + static const GVecGen3 ops[4] = { + { .fniv = gen_sabd_vec, + .fno = gen_helper_gvec_sabd_b, + .opt_opc = vecop_list, + .vece = MO_8 }, + { .fniv = gen_sabd_vec, + .fno = gen_helper_gvec_sabd_h, + .opt_opc = vecop_list, + .vece = MO_16 }, + { .fni4 = gen_sabd_i32, + .fniv = gen_sabd_vec, + .fno = gen_helper_gvec_sabd_s, + .opt_opc = vecop_list, + .vece = MO_32 }, + { .fni8 = gen_sabd_i64, + .fniv = gen_sabd_vec, + .fno = gen_helper_gvec_sabd_d, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .opt_opc = vecop_list, + .vece = MO_64 }, + }; + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); +} + +static void gen_uabd_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b) +{ + TCGv_i32 t = tcg_temp_new_i32(); + + tcg_gen_sub_i32(t, a, b); + tcg_gen_sub_i32(d, b, a); + tcg_gen_movcond_i32(TCG_COND_LTU, d, a, b, d, t); + tcg_temp_free_i32(t); +} + +static void gen_uabd_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b) +{ + TCGv_i64 t = tcg_temp_new_i64(); + + tcg_gen_sub_i64(t, a, b); + tcg_gen_sub_i64(d, b, a); + tcg_gen_movcond_i64(TCG_COND_LTU, d, a, b, d, t); + tcg_temp_free_i64(t); +} + +static void gen_uabd_vec(unsigned vece, TCGv_vec d, TCGv_vec a, TCGv_vec b) +{ + TCGv_vec t = tcg_temp_new_vec_matching(d); + + tcg_gen_umin_vec(vece, t, a, b); + tcg_gen_umax_vec(vece, d, a, b); + tcg_gen_sub_vec(vece, d, d, t); + tcg_temp_free_vec(t); +} + +void gen_gvec_uabd(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_sub_vec, INDEX_op_umin_vec, INDEX_op_umax_vec, 0 + }; + static const GVecGen3 ops[4] = { + { .fniv = gen_uabd_vec, + .fno = gen_helper_gvec_uabd_b, + .opt_opc = vecop_list, + .vece = MO_8 }, + { .fniv = gen_uabd_vec, + .fno = gen_helper_gvec_uabd_h, + .opt_opc = vecop_list, + .vece = MO_16 }, + { .fni4 = gen_uabd_i32, + .fniv = gen_uabd_vec, + .fno = gen_helper_gvec_uabd_s, + .opt_opc = vecop_list, + .vece = MO_32 }, + { .fni8 = gen_uabd_i64, + .fniv = gen_uabd_vec, + .fno = gen_helper_gvec_uabd_d, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .opt_opc = vecop_list, + .vece = MO_64 }, + }; + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); +} + /* Translate a NEON data processing instruction. Return nonzero if the instruction is invalid. We process data in a mixture of 32-bit and 64-bit chunks. @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) } return 1; + case NEON_3R_VABD: + if (u) { + gen_gvec_uabd(size, rd_ofs, rn_ofs, rm_ofs, + vec_size, vec_size); + } else { + gen_gvec_sabd(size, rd_ofs, rn_ofs, rm_ofs, + vec_size, vec_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_VQRSHL: GEN_NEON_INTEGER_OP_ENV(qrshl); break; - case NEON_3R_VABD: - GEN_NEON_INTEGER_OP(abd); - break; case NEON_3R_VABA: GEN_NEON_INTEGER_OP(abd); tcg_temp_free_i32(tmp2); diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vec_helper.c +++ b/target/arm/vec_helper.c @@ -XXX,XX +XXX,XX @@ DO_CMP0(gvec_cgt0_h, int16_t, >) DO_CMP0(gvec_cge0_h, int16_t, >=) #undef DO_CMP0 + +#define DO_ABD(NAME, TYPE) \ +void HELPER(NAME)(void *vd, void *vn, void *vm, uint32_t desc) \ +{ \ + intptr_t i, opr_sz = simd_oprsz(desc); \ + TYPE *d = vd, *n = vn, *m = vm; \ + \ + for (i = 0; i < opr_sz / sizeof(TYPE); ++i) { \ + d[i] = n[i] < m[i] ? m[i] - n[i] : n[i] - m[i]; \ + } \ + clear_tail(d, opr_sz, simd_maxsz(desc)); \ +} + +DO_ABD(gvec_sabd_b, int8_t) +DO_ABD(gvec_sabd_h, int16_t) +DO_ABD(gvec_sabd_s, int32_t) +DO_ABD(gvec_sabd_d, int64_t) + +DO_ABD(gvec_uabd_b, uint8_t) +DO_ABD(gvec_uabd_h, uint16_t) +DO_ABD(gvec_uabd_s, uint32_t) +DO_ABD(gvec_uabd_d, uint64_t) + +#undef DO_ABD -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Include 64-bit element size in preparation for SVE2. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200513163245.17915-17-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 17 +++-- target/arm/translate.h | 5 ++ target/arm/neon_helper.c | 10 --- target/arm/translate-a64.c | 17 ++--- target/arm/translate.c | 134 +++++++++++++++++++++++++++++++++++-- target/arm/vec_helper.c | 24 +++++++ 6 files changed, 174 insertions(+), 33 deletions(-) diff --git a/target/arm/helper.h b/target/arm/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_2(neon_pmax_s8, i32, i32, i32) DEF_HELPER_2(neon_pmax_u16, i32, i32, i32) DEF_HELPER_2(neon_pmax_s16, i32, i32, i32) -DEF_HELPER_2(neon_abd_u8, i32, i32, i32) -DEF_HELPER_2(neon_abd_s8, i32, i32, i32) -DEF_HELPER_2(neon_abd_u16, i32, i32, i32) -DEF_HELPER_2(neon_abd_s16, i32, i32, i32) -DEF_HELPER_2(neon_abd_u32, i32, i32, i32) -DEF_HELPER_2(neon_abd_s32, i32, i32, i32) - DEF_HELPER_2(neon_shl_u16, i32, i32, i32) DEF_HELPER_2(neon_shl_s16, i32, i32, i32) DEF_HELPER_2(neon_rshl_u8, i32, i32, i32) @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(gvec_uabd_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(gvec_uabd_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(gvec_uabd_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_saba_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_saba_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_saba_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_saba_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(gvec_uaba_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_uaba_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_uaba_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_uaba_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + #ifdef TARGET_AARCH64 #include "helper-a64.h" #include "helper-sve.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_gvec_sabd(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, void gen_gvec_uabd(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_saba(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); +void gen_gvec_uaba(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); + /* * Forward to the isar_feature_* tests given a DisasContext pointer. */ diff --git a/target/arm/neon_helper.c b/target/arm/neon_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon_helper.c +++ b/target/arm/neon_helper.c @@ -XXX,XX +XXX,XX @@ NEON_POP(pmax_s16, neon_s16, 2) NEON_POP(pmax_u16, neon_u16, 2) #undef NEON_FN -#define NEON_FN(dest, src1, src2) \ - dest = (src1 > src2) ? (src1 - src2) : (src2 - src1) -NEON_VOP(abd_s8, neon_s8, 4) -NEON_VOP(abd_u8, neon_u8, 4) -NEON_VOP(abd_s16, neon_s16, 2) -NEON_VOP(abd_u16, neon_u16, 2) -NEON_VOP(abd_s32, neon_s32, 1) -NEON_VOP(abd_u32, neon_u32, 1) -#undef NEON_FN - #define NEON_FN(dest, src1, src2) do { \ int8_t tmp; \ tmp = (int8_t)src2; \ diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_3same_int(DisasContext *s, uint32_t insn) gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_sabd, size); } return; + case 0xf: /* SABA, UABA */ + if (u) { + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_uaba, size); + } else { + gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_saba, size); + } + return; case 0x10: /* ADD, SUB */ if (u) { gen_gvec_fn3(s, is_q, rd, rn, rm, tcg_gen_gvec_sub, size); @@ -XXX,XX +XXX,XX @@ static void disas_simd_3same_int(DisasContext *s, uint32_t insn) genenvfn = fns[size][u]; break; } - case 0xf: /* SABA, UABA */ - { - static NeonGenTwoOpFn * const fns[3][2] = { - { gen_helper_neon_abd_s8, gen_helper_neon_abd_u8 }, - { gen_helper_neon_abd_s16, gen_helper_neon_abd_u16 }, - { gen_helper_neon_abd_s32, gen_helper_neon_abd_u32 }, - }; - genfn = fns[size][u]; - break; - } case 0x16: /* SQDMULH, SQRDMULH */ { static NeonGenTwoOpEnvFn * const fns[2][2] = { diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ void gen_gvec_uabd(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); } +static void gen_saba_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b) +{ + TCGv_i32 t = tcg_temp_new_i32(); + gen_sabd_i32(t, a, b); + tcg_gen_add_i32(d, d, t); + tcg_temp_free_i32(t); +} + +static void gen_saba_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b) +{ + TCGv_i64 t = tcg_temp_new_i64(); + gen_sabd_i64(t, a, b); + tcg_gen_add_i64(d, d, t); + tcg_temp_free_i64(t); +} + +static void gen_saba_vec(unsigned vece, TCGv_vec d, TCGv_vec a, TCGv_vec b) +{ + TCGv_vec t = tcg_temp_new_vec_matching(d); + gen_sabd_vec(vece, t, a, b); + tcg_gen_add_vec(vece, d, d, t); + tcg_temp_free_vec(t); +} + +void gen_gvec_saba(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_sub_vec, INDEX_op_add_vec, + INDEX_op_smin_vec, INDEX_op_smax_vec, 0 + }; + static const GVecGen3 ops[4] = { + { .fniv = gen_saba_vec, + .fno = gen_helper_gvec_saba_b, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_8 }, + { .fniv = gen_saba_vec, + .fno = gen_helper_gvec_saba_h, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_16 }, + { .fni4 = gen_saba_i32, + .fniv = gen_saba_vec, + .fno = gen_helper_gvec_saba_s, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_32 }, + { .fni8 = gen_saba_i64, + .fniv = gen_saba_vec, + .fno = gen_helper_gvec_saba_d, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_64 }, + }; + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); +} + +static void gen_uaba_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b) +{ + TCGv_i32 t = tcg_temp_new_i32(); + gen_uabd_i32(t, a, b); + tcg_gen_add_i32(d, d, t); + tcg_temp_free_i32(t); +} + +static void gen_uaba_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b) +{ + TCGv_i64 t = tcg_temp_new_i64(); + gen_uabd_i64(t, a, b); + tcg_gen_add_i64(d, d, t); + tcg_temp_free_i64(t); +} + +static void gen_uaba_vec(unsigned vece, TCGv_vec d, TCGv_vec a, TCGv_vec b) +{ + TCGv_vec t = tcg_temp_new_vec_matching(d); + gen_uabd_vec(vece, t, a, b); + tcg_gen_add_vec(vece, d, d, t); + tcg_temp_free_vec(t); +} + +void gen_gvec_uaba(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz) +{ + static const TCGOpcode vecop_list[] = { + INDEX_op_sub_vec, INDEX_op_add_vec, + INDEX_op_umin_vec, INDEX_op_umax_vec, 0 + }; + static const GVecGen3 ops[4] = { + { .fniv = gen_uaba_vec, + .fno = gen_helper_gvec_uaba_b, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_8 }, + { .fniv = gen_uaba_vec, + .fno = gen_helper_gvec_uaba_h, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_16 }, + { .fni4 = gen_uaba_i32, + .fniv = gen_uaba_vec, + .fno = gen_helper_gvec_uaba_s, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_32 }, + { .fni8 = gen_uaba_i64, + .fniv = gen_uaba_vec, + .fno = gen_helper_gvec_uaba_d, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + .opt_opc = vecop_list, + .load_dest = true, + .vece = MO_64 }, + }; + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, opr_sz, max_sz, &ops[vece]); +} + /* Translate a NEON data processing instruction. Return nonzero if the instruction is invalid. We process data in a mixture of 32-bit and 64-bit chunks. @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) } return 0; + case NEON_3R_VABA: + if (u) { + gen_gvec_uaba(size, rd_ofs, rn_ofs, rm_ofs, + vec_size, vec_size); + } else { + gen_gvec_saba(size, rd_ofs, rn_ofs, rm_ofs, + vec_size, vec_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_VQRSHL: GEN_NEON_INTEGER_OP_ENV(qrshl); break; - case NEON_3R_VABA: - GEN_NEON_INTEGER_OP(abd); - tcg_temp_free_i32(tmp2); - tmp2 = neon_load_reg(rd, pass); - gen_neon_add(size, tmp, tmp2); - break; case NEON_3R_VPMAX: GEN_NEON_INTEGER_OP(pmax); break; diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vec_helper.c +++ b/target/arm/vec_helper.c @@ -XXX,XX +XXX,XX @@ DO_ABD(gvec_uabd_s, uint32_t) DO_ABD(gvec_uabd_d, uint64_t) #undef DO_ABD + +#define DO_ABA(NAME, TYPE) \ +void HELPER(NAME)(void *vd, void *vn, void *vm, uint32_t desc) \ +{ \ + intptr_t i, opr_sz = simd_oprsz(desc); \ + TYPE *d = vd, *n = vn, *m = vm; \ + \ + for (i = 0; i < opr_sz / sizeof(TYPE); ++i) { \ + d[i] += n[i] < m[i] ? m[i] - n[i] : n[i] - m[i]; \ + } \ + clear_tail(d, opr_sz, simd_maxsz(desc)); \ +} + +DO_ABA(gvec_saba_b, int8_t) +DO_ABA(gvec_saba_h, int16_t) +DO_ABA(gvec_saba_s, int32_t) +DO_ABA(gvec_saba_d, int64_t) + +DO_ABA(gvec_uaba_b, uint8_t) +DO_ABA(gvec_uaba_h, uint16_t) +DO_ABA(gvec_uaba_s, uint32_t) +DO_ABA(gvec_uaba_d, uint64_t) + +#undef DO_ABA -- 2.20.1 From: Patrick Williams <patrick@stwcx.xyz> Sonora Pass is a 2 socket x86 motherboard designed by Facebook and supported by OpenBMC. Strapping configuration was obtained from hardware and i2c configuration is based on dts found at: https://github.com/facebook/openbmc-linux/blob/1633c87b8ba7c162095787c988979b748ba65dc8/arch/arm/boot/dts/aspeed-bmc-facebook-sonorapass.dts Booted a test image of http://github.com/facebook/openbmc to login prompt. Signed-off-by: Patrick Williams <patrick@stwcx.xyz> Reviewed-by: Amithash Prasad <amithash@fb.com> Reviewed-by: Cédric Le Goater <clg@kaod.org> [PMM: fixed block comment style nit] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/aspeed.c | 78 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 78 insertions(+) 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 @@ struct AspeedBoardState { SCU_AST2500_HW_STRAP_ACPI_ENABLE | \ SCU_HW_STRAP_SPI_MODE(SCU_HW_STRAP_SPI_MASTER)) +/* Sonorapass hardware value: 0xF100D216 */ +#define SONORAPASS_BMC_HW_STRAP1 ( \ + SCU_AST2500_HW_STRAP_SPI_AUTOFETCH_ENABLE | \ + SCU_AST2500_HW_STRAP_GPIO_STRAP_ENABLE | \ + SCU_AST2500_HW_STRAP_UART_DEBUG | \ + SCU_AST2500_HW_STRAP_RESERVED28 | \ + SCU_AST2500_HW_STRAP_DDR4_ENABLE | \ + SCU_HW_STRAP_VGA_CLASS_CODE | \ + SCU_HW_STRAP_LPC_RESET_PIN | \ + SCU_HW_STRAP_SPI_MODE(SCU_HW_STRAP_SPI_MASTER) | \ + SCU_AST2500_HW_STRAP_SET_AXI_AHB_RATIO(AXI_AHB_RATIO_2_1) | \ + SCU_HW_STRAP_VGA_BIOS_ROM | \ + SCU_HW_STRAP_VGA_SIZE_SET(VGA_16M_DRAM) | \ + SCU_AST2500_HW_STRAP_RESERVED1) + /* Swift hardware value: 0xF11AD206 */ #define SWIFT_BMC_HW_STRAP1 ( \ AST2500_HW_STRAP1_DEFAULTS | \ @@ -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) +{ + AspeedSoCState *soc = &bmc->soc; + + /* bus 2 : */ + i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 2), "tmp105", 0x48); + i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 2), "tmp105", 0x49); + /* bus 2 : pca9546 @ 0x73 */ + + /* bus 3 : pca9548 @ 0x70 */ + + /* bus 4 : */ + uint8_t *eeprom4_54 = g_malloc0(8 * 1024); + smbus_eeprom_init_one(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 4), 0x54, + eeprom4_54); + /* PCA9539 @ 0x76, but PCA9552 is compatible */ + i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 4), "pca9552", 0x76); + /* PCA9539 @ 0x77, but PCA9552 is compatible */ + i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 4), "pca9552", 0x77); + + /* bus 6 : */ + i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 6), "tmp105", 0x48); + i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 6), "tmp105", 0x49); + /* bus 6 : pca9546 @ 0x73 */ + + /* bus 8 : */ + uint8_t *eeprom8_56 = g_malloc0(8 * 1024); + smbus_eeprom_init_one(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 8), 0x56, + eeprom8_56); + i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 8), "pca9552", 0x60); + i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 8), "pca9552", 0x61); + /* bus 8 : adc128d818 @ 0x1d */ + /* bus 8 : adc128d818 @ 0x1f */ + + /* + * bus 13 : pca9548 @ 0x71 + * - channel 3: + * - tmm421 @ 0x4c + * - tmp421 @ 0x4e + * - tmp421 @ 0x4f + */ + +} + static void witherspoon_bmc_i2c_init(AspeedBoardState *bmc) { AspeedSoCState *soc = &bmc->soc; @@ -XXX,XX +XXX,XX @@ static void aspeed_machine_romulus_class_init(ObjectClass *oc, void *data) mc->default_ram_size = 512 * MiB; }; +static void aspeed_machine_sonorapass_class_init(ObjectClass *oc, void *data) +{ + MachineClass *mc = MACHINE_CLASS(oc); + AspeedMachineClass *amc = ASPEED_MACHINE_CLASS(oc); + + mc->desc = "OCP SonoraPass BMC (ARM1176)"; + amc->soc_name = "ast2500-a1"; + amc->hw_strap1 = SONORAPASS_BMC_HW_STRAP1; + amc->fmc_model = "mx66l1g45g"; + amc->spi_model = "mx66l1g45g"; + amc->num_cs = 2; + amc->i2c_init = sonorapass_bmc_i2c_init; + mc->default_ram_size = 512 * MiB; +}; + static void aspeed_machine_swift_class_init(ObjectClass *oc, void *data) { MachineClass *mc = MACHINE_CLASS(oc); @@ -XXX,XX +XXX,XX @@ static const TypeInfo aspeed_machine_types[] = { .name = MACHINE_TYPE_NAME("swift-bmc"), .parent = TYPE_ASPEED_MACHINE, .class_init = aspeed_machine_swift_class_init, + }, { + .name = MACHINE_TYPE_NAME("sonorapass-bmc"), + .parent = TYPE_ASPEED_MACHINE, + .class_init = aspeed_machine_sonorapass_class_init, }, { .name = MACHINE_TYPE_NAME("witherspoon-bmc"), .parent = TYPE_ASPEED_MACHINE, -- 2.20.1 From: Dongjiu Geng <gengdongjiu@huawei.com> The little end UUID is used in many places, so make NVDIMM_UUID_LE to a common macro to convert the UUID to a little end array. Reviewed-by: Xiang Zheng <zhengxiang9@huawei.com> Signed-off-by: Dongjiu Geng <gengdongjiu@huawei.com> Message-id: 20200512030609.19593-2-gengdongjiu@huawei.com Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/qemu/uuid.h | 27 +++++++++++++++++++++++++++ hw/acpi/nvdimm.c | 10 +++------- 2 files changed, 30 insertions(+), 7 deletions(-) diff --git a/include/qemu/uuid.h b/include/qemu/uuid.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/uuid.h +++ b/include/qemu/uuid.h @@ -XXX,XX +XXX,XX @@ typedef struct { }; } QemuUUID; +/** + * UUID_LE - converts the fields of UUID to little-endian array, + * each of parameters is the filed of UUID. + * + * @time_low: The low field of the timestamp + * @time_mid: The middle field of the timestamp + * @time_hi_and_version: The high field of the timestamp + * multiplexed with the version number + * @clock_seq_hi_and_reserved: The high field of the clock + * sequence multiplexed with the variant + * @clock_seq_low: The low field of the clock sequence + * @node0: The spatially unique node0 identifier + * @node1: The spatially unique node1 identifier + * @node2: The spatially unique node2 identifier + * @node3: The spatially unique node3 identifier + * @node4: The spatially unique node4 identifier + * @node5: The spatially unique node5 identifier + */ +#define UUID_LE(time_low, time_mid, time_hi_and_version, \ + clock_seq_hi_and_reserved, clock_seq_low, node0, node1, node2, \ + node3, node4, node5) \ + { (time_low) & 0xff, ((time_low) >> 8) & 0xff, ((time_low) >> 16) & 0xff, \ + ((time_low) >> 24) & 0xff, (time_mid) & 0xff, ((time_mid) >> 8) & 0xff, \ + (time_hi_and_version) & 0xff, ((time_hi_and_version) >> 8) & 0xff, \ + (clock_seq_hi_and_reserved), (clock_seq_low), (node0), (node1), (node2),\ + (node3), (node4), (node5) } + #define UUID_FMT "%02hhx%02hhx%02hhx%02hhx-" \ "%02hhx%02hhx-%02hhx%02hhx-" \ "%02hhx%02hhx-" \ diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c index XXXXXXX..XXXXXXX 100644 --- a/hw/acpi/nvdimm.c +++ b/hw/acpi/nvdimm.c @@ -XXX,XX +XXX,XX @@ */ #include "qemu/osdep.h" +#include "qemu/uuid.h" #include "hw/acpi/acpi.h" #include "hw/acpi/aml-build.h" #include "hw/acpi/bios-linker-loader.h" @@ -XXX,XX +XXX,XX @@ #include "hw/mem/nvdimm.h" #include "qemu/nvdimm-utils.h" -#define NVDIMM_UUID_LE(a, b, c, d0, d1, d2, d3, d4, d5, d6, d7) \ - { (a) & 0xff, ((a) >> 8) & 0xff, ((a) >> 16) & 0xff, ((a) >> 24) & 0xff, \ - (b) & 0xff, ((b) >> 8) & 0xff, (c) & 0xff, ((c) >> 8) & 0xff, \ - (d0), (d1), (d2), (d3), (d4), (d5), (d6), (d7) } - /* * define Byte Addressable Persistent Memory (PM) Region according to * ACPI 6.0: 5.2.25.1 System Physical Address Range Structure. */ static const uint8_t nvdimm_nfit_spa_uuid[] = - NVDIMM_UUID_LE(0x66f0d379, 0xb4f3, 0x4074, 0xac, 0x43, 0x0d, 0x33, - 0x18, 0xb7, 0x8c, 0xdb); + UUID_LE(0x66f0d379, 0xb4f3, 0x4074, 0xac, 0x43, 0x0d, 0x33, + 0x18, 0xb7, 0x8c, 0xdb); /* * NVDIMM Firmware Interface Table -- 2.20.1 From: Dongjiu Geng <gengdongjiu@huawei.com> RAS Virtualization feature is not supported now, so add a RAS machine option and disable it by default. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Dongjiu Geng <gengdongjiu@huawei.com> Signed-off-by: Xiang Zheng <zhengxiang9@huawei.com> Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Reviewed-by: Igor Mammedov <imammedo@redhat.com> Message-id: 20200512030609.19593-3-gengdongjiu@huawei.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/virt.h | 1 + hw/arm/virt.c | 23 +++++++++++++++++++++++ 2 files changed, 24 insertions(+) diff --git a/include/hw/arm/virt.h b/include/hw/arm/virt.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/virt.h +++ b/include/hw/arm/virt.h @@ -XXX,XX +XXX,XX @@ typedef struct { bool highmem_ecam; bool its; bool virt; + bool ras; OnOffAuto acpi; VirtGICType gic_version; VirtIOMMUType iommu; diff --git a/hw/arm/virt.c b/hw/arm/virt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/virt.c +++ b/hw/arm/virt.c @@ -XXX,XX +XXX,XX @@ static void virt_set_acpi(Object *obj, Visitor *v, const char *name, visit_type_OnOffAuto(v, name, &vms->acpi, errp); } +static bool virt_get_ras(Object *obj, Error **errp) +{ + VirtMachineState *vms = VIRT_MACHINE(obj); + + return vms->ras; +} + +static void virt_set_ras(Object *obj, bool value, Error **errp) +{ + VirtMachineState *vms = VIRT_MACHINE(obj); + + vms->ras = value; +} + static char *virt_get_gic_version(Object *obj, Error **errp) { VirtMachineState *vms = VIRT_MACHINE(obj); @@ -XXX,XX +XXX,XX @@ static void virt_instance_init(Object *obj) "Valid values are none and smmuv3", NULL); + /* Default disallows RAS instantiation */ + vms->ras = false; + object_property_add_bool(obj, "ras", virt_get_ras, + virt_set_ras, NULL); + object_property_set_description(obj, "ras", + "Set on/off to enable/disable reporting host memory errors " + "to a KVM guest using ACPI and guest external abort exceptions", + NULL); + vms->irqmap = a15irqmap; virt_flash_create(vms); -- 2.20.1 From: Dongjiu Geng <gengdongjiu@huawei.com> Add APEI/GHES detailed design document Signed-off-by: Dongjiu Geng <gengdongjiu@huawei.com> Signed-off-by: Xiang Zheng <zhengxiang9@huawei.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Reviewed-by: Igor Mammedov <imammedo@redhat.com> Message-id: 20200512030609.19593-4-gengdongjiu@huawei.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/specs/acpi_hest_ghes.rst | 110 ++++++++++++++++++++++++++++++++++ docs/specs/index.rst | 1 + 2 files changed, 111 insertions(+) create mode 100644 docs/specs/acpi_hest_ghes.rst diff --git a/docs/specs/acpi_hest_ghes.rst b/docs/specs/acpi_hest_ghes.rst new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/docs/specs/acpi_hest_ghes.rst @@ -XXX,XX +XXX,XX @@ +APEI tables generating and CPER record +====================================== + +.. + Copyright (c) 2020 HUAWEI TECHNOLOGIES CO., LTD. + + This work is licensed under the terms of the GNU GPL, version 2 or later. + See the COPYING file in the top-level directory. + +Design Details +-------------- + +:: + + etc/acpi/tables etc/hardware_errors + ==================== =============================== + + +--------------------------+ +----------------------------+ + | | HEST | +--------->| error_block_address1 |------+ + | +--------------------------+ | +----------------------------+ | + | | GHES1 | | +------->| error_block_address2 |------+-+ + | +--------------------------+ | | +----------------------------+ | | + | | ................. | | | | .............. | | | + | | error_status_address-----+-+ | -----------------------------+ | | + | | ................. | | +--->| error_block_addressN |------+-+---+ + | | read_ack_register--------+-+ | | +----------------------------+ | | | + | | read_ack_preserve | +-+---+--->| read_ack_register1 | | | | + | | read_ack_write | | | +----------------------------+ | | | + + +--------------------------+ | +-+--->| read_ack_register2 | | | | + | | GHES2 | | | | +----------------------------+ | | | + + +--------------------------+ | | | | ............. | | | | + | | ................. | | | | +----------------------------+ | | | + | | error_status_address-----+---+ | | +->| read_ack_registerN | | | | + | | ................. | | | | +----------------------------+ | | | + | | read_ack_register--------+-----+ | | |Generic Error Status Block 1|<-----+ | | + | | read_ack_preserve | | | |-+------------------------+-+ | | + | | read_ack_write | | | | | CPER | | | | + + +--------------------------| | | | | CPER | | | | + | | ............... | | | | | .... | | | | + + +--------------------------+ | | | | CPER | | | | + | | GHESN | | | |-+------------------------+-| | | + + +--------------------------+ | | |Generic Error Status Block 2|<-------+ | + | | ................. | | | |-+------------------------+-+ | + | | error_status_address-----+-------+ | | | CPER | | | + | | ................. | | | | CPER | | | + | | read_ack_register--------+---------+ | | .... | | | + | | read_ack_preserve | | | CPER | | | + | | read_ack_write | +-+------------------------+-+ | + + +--------------------------+ | .......... | | + |----------------------------+ | + |Generic Error Status Block N |<----------+ + |-+-------------------------+-+ + | | CPER | | + | | CPER | | + | | .... | | + | | CPER | | + +-+-------------------------+-+ + + +(1) QEMU generates the ACPI HEST table. This table goes in the current + "etc/acpi/tables" fw_cfg blob. Each error source has different + notification types. + +(2) A new fw_cfg blob called "etc/hardware_errors" is introduced. QEMU + also needs to populate this blob. The "etc/hardware_errors" fw_cfg blob + contains an address registers table and an Error Status Data Block table. + +(3) The address registers table contains N Error Block Address entries + and N Read Ack Register entries. The size for each entry is 8-byte. + The Error Status Data Block table contains N Error Status Data Block + entries. The size for each entry is 4096(0x1000) bytes. The total size + for the "etc/hardware_errors" fw_cfg blob is (N * 8 * 2 + N * 4096) bytes. + N is the number of the kinds of hardware error sources. + +(4) QEMU generates the ACPI linker/loader script for the firmware. The + firmware pre-allocates memory for "etc/acpi/tables", "etc/hardware_errors" + and copies blob contents there. + +(5) QEMU generates N ADD_POINTER commands, which patch addresses in the + "error_status_address" fields of the HEST table with a pointer to the + corresponding "address registers" in the "etc/hardware_errors" blob. + +(6) QEMU generates N ADD_POINTER commands, which patch addresses in the + "read_ack_register" fields of the HEST table with a pointer to the + corresponding "read_ack_register" within the "etc/hardware_errors" blob. + +(7) QEMU generates N ADD_POINTER commands for the firmware, which patch + addresses in the "error_block_address" fields with a pointer to the + respective "Error Status Data Block" in the "etc/hardware_errors" blob. + +(8) QEMU defines a third and write-only fw_cfg blob which is called + "etc/hardware_errors_addr". Through that blob, the firmware can send back + the guest-side allocation addresses to QEMU. The "etc/hardware_errors_addr" + blob contains a 8-byte entry. QEMU generates a single WRITE_POINTER command + for the firmware. The firmware will write back the start address of + "etc/hardware_errors" blob to the fw_cfg file "etc/hardware_errors_addr". + +(9) When QEMU gets a SIGBUS from the kernel, QEMU writes CPER into corresponding + "Error Status Data Block", guest memory, and then injects platform specific + interrupt (in case of arm/virt machine it's Synchronous External Abort) as a + notification which is necessary for notifying the guest. + +(10) This notification (in virtual hardware) will be handled by the guest + kernel, on receiving notification, guest APEI driver could read the CPER error + and take appropriate action. + +(11) kvm_arch_on_sigbus_vcpu() uses source_id as index in "etc/hardware_errors" to + find out "Error Status Data Block" entry corresponding to error source. So supported + source_id values should be assigned here and not be changed afterwards to make sure + that guest will write error into expected "Error Status Data Block" even if guest was + migrated to a newer QEMU. diff --git a/docs/specs/index.rst b/docs/specs/index.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/specs/index.rst +++ b/docs/specs/index.rst @@ -XXX,XX +XXX,XX @@ Contents: ppc-spapr-xive acpi_hw_reduced_hotplug tpm + acpi_hest_ghes -- 2.20.1 From: Dongjiu Geng <gengdongjiu@huawei.com> This patch builds error_block_address and read_ack_register fields in hardware errors table , the error_block_address points to Generic Error Status Block(GESB) via bios_linker. The max size for one GESB is 1kb, For more detailed information, please refer to document: docs/specs/acpi_hest_ghes.rst Now we only support one Error source, if necessary, we can extend to support more. Suggested-by: Laszlo Ersek <lersek@redhat.com> Signed-off-by: Xiang Zheng <zhengxiang9@huawei.com> Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Reviewed-by: Igor Mammedov <imammedo@redhat.com> Signed-off-by: Dongjiu Geng <gengdongjiu@huawei.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Message-id: 20200512030609.19593-5-gengdongjiu@huawei.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- default-configs/arm-softmmu.mak | 1 + include/hw/acpi/aml-build.h | 1 + include/hw/acpi/ghes.h | 28 +++++++++++ hw/acpi/aml-build.c | 2 + hw/acpi/ghes.c | 89 +++++++++++++++++++++++++++++++++ hw/arm/virt-acpi-build.c | 5 ++ hw/acpi/Kconfig | 4 ++ hw/acpi/Makefile.objs | 1 + 8 files changed, 131 insertions(+) create mode 100644 include/hw/acpi/ghes.h create mode 100644 hw/acpi/ghes.c diff --git a/default-configs/arm-softmmu.mak b/default-configs/arm-softmmu.mak index XXXXXXX..XXXXXXX 100644 --- a/default-configs/arm-softmmu.mak +++ b/default-configs/arm-softmmu.mak @@ -XXX,XX +XXX,XX @@ CONFIG_FSL_IMX7=y CONFIG_FSL_IMX6UL=y CONFIG_SEMIHOSTING=y CONFIG_ALLWINNER_H3=y +CONFIG_ACPI_APEI=y diff --git a/include/hw/acpi/aml-build.h b/include/hw/acpi/aml-build.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/acpi/aml-build.h +++ b/include/hw/acpi/aml-build.h @@ -XXX,XX +XXX,XX @@ struct AcpiBuildTables { GArray *rsdp; GArray *tcpalog; GArray *vmgenid; + GArray *hardware_errors; BIOSLinker *linker; } AcpiBuildTables; diff --git a/include/hw/acpi/ghes.h b/include/hw/acpi/ghes.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/acpi/ghes.h @@ -XXX,XX +XXX,XX @@ +/* + * Support for generating APEI tables and recording CPER for Guests + * + * Copyright (c) 2020 HUAWEI TECHNOLOGIES CO., LTD. + * + * Author: Dongjiu Geng <gengdongjiu@huawei.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + + * This program 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 General Public License for more details. + + * You should have received a copy of the GNU General Public License along + * with this program; if not, see <http://www.gnu.org/licenses/>. + */ + +#ifndef ACPI_GHES_H +#define ACPI_GHES_H + +#include "hw/acpi/bios-linker-loader.h" + +void build_ghes_error_table(GArray *hardware_errors, BIOSLinker *linker); +#endif diff --git a/hw/acpi/aml-build.c b/hw/acpi/aml-build.c index XXXXXXX..XXXXXXX 100644 --- a/hw/acpi/aml-build.c +++ b/hw/acpi/aml-build.c @@ -XXX,XX +XXX,XX @@ void acpi_build_tables_init(AcpiBuildTables *tables) tables->table_data = g_array_new(false, true /* clear */, 1); tables->tcpalog = g_array_new(false, true /* clear */, 1); tables->vmgenid = g_array_new(false, true /* clear */, 1); + tables->hardware_errors = g_array_new(false, true /* clear */, 1); tables->linker = bios_linker_loader_init(); } @@ -XXX,XX +XXX,XX @@ void acpi_build_tables_cleanup(AcpiBuildTables *tables, bool mfre) g_array_free(tables->table_data, true); g_array_free(tables->tcpalog, mfre); g_array_free(tables->vmgenid, mfre); + g_array_free(tables->hardware_errors, mfre); } /* diff --git a/hw/acpi/ghes.c b/hw/acpi/ghes.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/acpi/ghes.c @@ -XXX,XX +XXX,XX @@ +/* + * Support for generating APEI tables and recording CPER for Guests + * + * Copyright (c) 2020 HUAWEI TECHNOLOGIES CO., LTD. + * + * Author: Dongjiu Geng <gengdongjiu@huawei.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + + * This program 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 General Public License for more details. + + * You should have received a copy of the GNU General Public License along + * with this program; if not, see <http://www.gnu.org/licenses/>. + */ + +#include "qemu/osdep.h" +#include "qemu/units.h" +#include "hw/acpi/ghes.h" +#include "hw/acpi/aml-build.h" + +#define ACPI_GHES_ERRORS_FW_CFG_FILE "etc/hardware_errors" +#define ACPI_GHES_DATA_ADDR_FW_CFG_FILE "etc/hardware_errors_addr" + +/* The max size in bytes for one error block */ +#define ACPI_GHES_MAX_RAW_DATA_LENGTH (1 * KiB) + +/* Now only support ARMv8 SEA notification type error source */ +#define ACPI_GHES_ERROR_SOURCE_COUNT 1 + +/* + * Build table for the hardware error fw_cfg blob. + * Initialize "etc/hardware_errors" and "etc/hardware_errors_addr" fw_cfg blobs. + * See docs/specs/acpi_hest_ghes.rst for blobs format. + */ +void build_ghes_error_table(GArray *hardware_errors, BIOSLinker *linker) +{ + int i, error_status_block_offset; + + /* Build error_block_address */ + for (i = 0; i < ACPI_GHES_ERROR_SOURCE_COUNT; i++) { + build_append_int_noprefix(hardware_errors, 0, sizeof(uint64_t)); + } + + /* Build read_ack_register */ + for (i = 0; i < ACPI_GHES_ERROR_SOURCE_COUNT; i++) { + /* + * Initialize the value of read_ack_register to 1, so GHES can be + * writeable after (re)boot. + * ACPI 6.2: 18.3.2.8 Generic Hardware Error Source version 2 + * (GHESv2 - Type 10) + */ + build_append_int_noprefix(hardware_errors, 1, sizeof(uint64_t)); + } + + /* Generic Error Status Block offset in the hardware error fw_cfg blob */ + error_status_block_offset = hardware_errors->len; + + /* Reserve space for Error Status Data Block */ + acpi_data_push(hardware_errors, + ACPI_GHES_MAX_RAW_DATA_LENGTH * ACPI_GHES_ERROR_SOURCE_COUNT); + + /* Tell guest firmware to place hardware_errors blob into RAM */ + bios_linker_loader_alloc(linker, ACPI_GHES_ERRORS_FW_CFG_FILE, + hardware_errors, sizeof(uint64_t), false); + + for (i = 0; i < ACPI_GHES_ERROR_SOURCE_COUNT; i++) { + /* + * Tell firmware to patch error_block_address entries to point to + * corresponding "Generic Error Status Block" + */ + bios_linker_loader_add_pointer(linker, + ACPI_GHES_ERRORS_FW_CFG_FILE, sizeof(uint64_t) * i, + sizeof(uint64_t), ACPI_GHES_ERRORS_FW_CFG_FILE, + error_status_block_offset + i * ACPI_GHES_MAX_RAW_DATA_LENGTH); + } + + /* + * tell firmware to write hardware_errors GPA into + * hardware_errors_addr fw_cfg, once the former has been initialized. + */ + bios_linker_loader_write_pointer(linker, ACPI_GHES_DATA_ADDR_FW_CFG_FILE, + 0, sizeof(uint64_t), ACPI_GHES_ERRORS_FW_CFG_FILE, 0); +} diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/virt-acpi-build.c +++ b/hw/arm/virt-acpi-build.c @@ -XXX,XX +XXX,XX @@ #include "sysemu/reset.h" #include "kvm_arm.h" #include "migration/vmstate.h" +#include "hw/acpi/ghes.h" #define ARM_SPI_BASE 32 @@ -XXX,XX +XXX,XX @@ void virt_acpi_build(VirtMachineState *vms, AcpiBuildTables *tables) acpi_add_table(table_offsets, tables_blob); build_spcr(tables_blob, tables->linker, vms); + if (vms->ras) { + build_ghes_error_table(tables->hardware_errors, tables->linker); + } + if (ms->numa_state->num_nodes > 0) { acpi_add_table(table_offsets, tables_blob); build_srat(tables_blob, tables->linker, vms); diff --git a/hw/acpi/Kconfig b/hw/acpi/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/acpi/Kconfig +++ b/hw/acpi/Kconfig @@ -XXX,XX +XXX,XX @@ config ACPI_HMAT bool depends on ACPI +config ACPI_APEI + bool + depends on ACPI + config ACPI_PCI bool depends on ACPI && PCI diff --git a/hw/acpi/Makefile.objs b/hw/acpi/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/hw/acpi/Makefile.objs +++ b/hw/acpi/Makefile.objs @@ -XXX,XX +XXX,XX @@ common-obj-$(CONFIG_ACPI_NVDIMM) += nvdimm.o common-obj-$(CONFIG_ACPI_VMGENID) += vmgenid.o common-obj-$(CONFIG_ACPI_HW_REDUCED) += generic_event_device.o common-obj-$(CONFIG_ACPI_HMAT) += hmat.o +common-obj-$(CONFIG_ACPI_APEI) += ghes.o common-obj-$(call lnot,$(CONFIG_ACPI_X86)) += acpi-stub.o common-obj-$(call lnot,$(CONFIG_PC)) += acpi-x86-stub.o -- 2.20.1 From: Dongjiu Geng <gengdongjiu@huawei.com> This patch builds Hardware Error Source Table(HEST) via fw_cfg blobs. Now it only supports ARMv8 SEA, a type of Generic Hardware Error Source version 2(GHESv2) error source. Afterwards, we can extend the supported types if needed. For the CPER section, currently it is memory section because kernel mainly wants userspace to handle the memory errors. This patch follows the spec ACPI 6.2 to build the Hardware Error Source table. For more detailed information, please refer to document: docs/specs/acpi_hest_ghes.rst build_ghes_hw_error_notification() helper will help to add Hardware Error Notification to ACPI tables without using packed C structures and avoid endianness issues as API doesn't need explicit conversion. Signed-off-by: Xiang Zheng <zhengxiang9@huawei.com> Signed-off-by: Dongjiu Geng <gengdongjiu@huawei.com> Reviewed-by: Igor Mammedov <imammedo@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Message-id: 20200512030609.19593-6-gengdongjiu@huawei.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/acpi/ghes.h | 39 ++++++++++++ hw/acpi/ghes.c | 126 +++++++++++++++++++++++++++++++++++++++ hw/arm/virt-acpi-build.c | 2 + 3 files changed, 167 insertions(+) diff --git a/include/hw/acpi/ghes.h b/include/hw/acpi/ghes.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/acpi/ghes.h +++ b/include/hw/acpi/ghes.h @@ -XXX,XX +XXX,XX @@ #include "hw/acpi/bios-linker-loader.h" +/* + * Values for Hardware Error Notification Type field + */ +enum AcpiGhesNotifyType { + /* Polled */ + ACPI_GHES_NOTIFY_POLLED = 0, + /* External Interrupt */ + ACPI_GHES_NOTIFY_EXTERNAL = 1, + /* Local Interrupt */ + ACPI_GHES_NOTIFY_LOCAL = 2, + /* SCI */ + ACPI_GHES_NOTIFY_SCI = 3, + /* NMI */ + ACPI_GHES_NOTIFY_NMI = 4, + /* CMCI, ACPI 5.0: 18.3.2.7, Table 18-290 */ + ACPI_GHES_NOTIFY_CMCI = 5, + /* MCE, ACPI 5.0: 18.3.2.7, Table 18-290 */ + ACPI_GHES_NOTIFY_MCE = 6, + /* GPIO-Signal, ACPI 6.0: 18.3.2.7, Table 18-332 */ + ACPI_GHES_NOTIFY_GPIO = 7, + /* ARMv8 SEA, ACPI 6.1: 18.3.2.9, Table 18-345 */ + ACPI_GHES_NOTIFY_SEA = 8, + /* ARMv8 SEI, ACPI 6.1: 18.3.2.9, Table 18-345 */ + ACPI_GHES_NOTIFY_SEI = 9, + /* External Interrupt - GSIV, ACPI 6.1: 18.3.2.9, Table 18-345 */ + ACPI_GHES_NOTIFY_GSIV = 10, + /* Software Delegated Exception, ACPI 6.2: 18.3.2.9, Table 18-383 */ + ACPI_GHES_NOTIFY_SDEI = 11, + /* 12 and greater are reserved */ + ACPI_GHES_NOTIFY_RESERVED = 12 +}; + +enum { + ACPI_HEST_SRC_ID_SEA = 0, + /* future ids go here */ + ACPI_HEST_SRC_ID_RESERVED, +}; + void build_ghes_error_table(GArray *hardware_errors, BIOSLinker *linker); +void acpi_build_hest(GArray *table_data, BIOSLinker *linker); #endif diff --git a/hw/acpi/ghes.c b/hw/acpi/ghes.c index XXXXXXX..XXXXXXX 100644 --- a/hw/acpi/ghes.c +++ b/hw/acpi/ghes.c @@ -XXX,XX +XXX,XX @@ #include "qemu/units.h" #include "hw/acpi/ghes.h" #include "hw/acpi/aml-build.h" +#include "qemu/error-report.h" #define ACPI_GHES_ERRORS_FW_CFG_FILE "etc/hardware_errors" #define ACPI_GHES_DATA_ADDR_FW_CFG_FILE "etc/hardware_errors_addr" @@ -XXX,XX +XXX,XX @@ /* Now only support ARMv8 SEA notification type error source */ #define ACPI_GHES_ERROR_SOURCE_COUNT 1 +/* Generic Hardware Error Source version 2 */ +#define ACPI_GHES_SOURCE_GENERIC_ERROR_V2 10 + +/* Address offset in Generic Address Structure(GAS) */ +#define GAS_ADDR_OFFSET 4 + +/* + * Hardware Error Notification + * ACPI 4.0: 17.3.2.7 Hardware Error Notification + * Composes dummy Hardware Error Notification descriptor of specified type + */ +static void build_ghes_hw_error_notification(GArray *table, const uint8_t type) +{ + /* Type */ + build_append_int_noprefix(table, type, 1); + /* + * Length: + * Total length of the structure in bytes + */ + build_append_int_noprefix(table, 28, 1); + /* Configuration Write Enable */ + build_append_int_noprefix(table, 0, 2); + /* Poll Interval */ + build_append_int_noprefix(table, 0, 4); + /* Vector */ + build_append_int_noprefix(table, 0, 4); + /* Switch To Polling Threshold Value */ + build_append_int_noprefix(table, 0, 4); + /* Switch To Polling Threshold Window */ + build_append_int_noprefix(table, 0, 4); + /* Error Threshold Value */ + build_append_int_noprefix(table, 0, 4); + /* Error Threshold Window */ + build_append_int_noprefix(table, 0, 4); +} + /* * Build table for the hardware error fw_cfg blob. * Initialize "etc/hardware_errors" and "etc/hardware_errors_addr" fw_cfg blobs. @@ -XXX,XX +XXX,XX @@ void build_ghes_error_table(GArray *hardware_errors, BIOSLinker *linker) bios_linker_loader_write_pointer(linker, ACPI_GHES_DATA_ADDR_FW_CFG_FILE, 0, sizeof(uint64_t), ACPI_GHES_ERRORS_FW_CFG_FILE, 0); } + +/* Build Generic Hardware Error Source version 2 (GHESv2) */ +static void build_ghes_v2(GArray *table_data, int source_id, BIOSLinker *linker) +{ + uint64_t address_offset; + /* + * Type: + * Generic Hardware Error Source version 2(GHESv2 - Type 10) + */ + build_append_int_noprefix(table_data, ACPI_GHES_SOURCE_GENERIC_ERROR_V2, 2); + /* Source Id */ + build_append_int_noprefix(table_data, source_id, 2); + /* Related Source Id */ + build_append_int_noprefix(table_data, 0xffff, 2); + /* Flags */ + build_append_int_noprefix(table_data, 0, 1); + /* Enabled */ + build_append_int_noprefix(table_data, 1, 1); + + /* Number of Records To Pre-allocate */ + build_append_int_noprefix(table_data, 1, 4); + /* Max Sections Per Record */ + build_append_int_noprefix(table_data, 1, 4); + /* Max Raw Data Length */ + build_append_int_noprefix(table_data, ACPI_GHES_MAX_RAW_DATA_LENGTH, 4); + + address_offset = table_data->len; + /* Error Status Address */ + build_append_gas(table_data, AML_AS_SYSTEM_MEMORY, 0x40, 0, + 4 /* QWord access */, 0); + bios_linker_loader_add_pointer(linker, ACPI_BUILD_TABLE_FILE, + address_offset + GAS_ADDR_OFFSET, sizeof(uint64_t), + ACPI_GHES_ERRORS_FW_CFG_FILE, source_id * sizeof(uint64_t)); + + switch (source_id) { + case ACPI_HEST_SRC_ID_SEA: + /* + * Notification Structure + * Now only enable ARMv8 SEA notification type + */ + build_ghes_hw_error_notification(table_data, ACPI_GHES_NOTIFY_SEA); + break; + default: + error_report("Not support this error source"); + abort(); + } + + /* Error Status Block Length */ + build_append_int_noprefix(table_data, ACPI_GHES_MAX_RAW_DATA_LENGTH, 4); + + /* + * Read Ack Register + * ACPI 6.1: 18.3.2.8 Generic Hardware Error Source + * version 2 (GHESv2 - Type 10) + */ + address_offset = table_data->len; + build_append_gas(table_data, AML_AS_SYSTEM_MEMORY, 0x40, 0, + 4 /* QWord access */, 0); + bios_linker_loader_add_pointer(linker, ACPI_BUILD_TABLE_FILE, + address_offset + GAS_ADDR_OFFSET, + sizeof(uint64_t), ACPI_GHES_ERRORS_FW_CFG_FILE, + (ACPI_GHES_ERROR_SOURCE_COUNT + source_id) * sizeof(uint64_t)); + + /* + * Read Ack Preserve field + * We only provide the first bit in Read Ack Register to OSPM to write + * while the other bits are preserved. + */ + build_append_int_noprefix(table_data, ~0x1ULL, 8); + /* Read Ack Write */ + build_append_int_noprefix(table_data, 0x1, 8); +} + +/* Build Hardware Error Source Table */ +void acpi_build_hest(GArray *table_data, BIOSLinker *linker) +{ + uint64_t hest_start = table_data->len; + + /* Hardware Error Source Table header*/ + acpi_data_push(table_data, sizeof(AcpiTableHeader)); + + /* Error Source Count */ + build_append_int_noprefix(table_data, ACPI_GHES_ERROR_SOURCE_COUNT, 4); + + build_ghes_v2(table_data, ACPI_HEST_SRC_ID_SEA, linker); + + build_header(linker, table_data, (void *)(table_data->data + hest_start), + "HEST", table_data->len - hest_start, 1, NULL, NULL); +} diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/virt-acpi-build.c +++ b/hw/arm/virt-acpi-build.c @@ -XXX,XX +XXX,XX @@ void virt_acpi_build(VirtMachineState *vms, AcpiBuildTables *tables) if (vms->ras) { build_ghes_error_table(tables->hardware_errors, tables->linker); + acpi_add_table(table_offsets, tables_blob); + acpi_build_hest(tables_blob, tables->linker); } if (ms->numa_state->num_nodes > 0) { -- 2.20.1 From: Dongjiu Geng <gengdongjiu@huawei.com> Record the GHEB address via fw_cfg file, when recording a error to CPER, it will use this address to find out Generic Error Data Entries and write the error. In order to avoid migration failure, make hardware error table address to a part of GED device instead of global variable, then this address will be migrated to target QEMU. Acked-by: Xiang Zheng <zhengxiang9@huawei.com> Signed-off-by: Dongjiu Geng <gengdongjiu@huawei.com> Reviewed-by: Igor Mammedov <imammedo@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Message-id: 20200512030609.19593-7-gengdongjiu@huawei.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/acpi/generic_event_device.h | 2 ++ include/hw/acpi/ghes.h | 6 ++++++ hw/acpi/generic_event_device.c | 19 +++++++++++++++++++ hw/acpi/ghes.c | 14 ++++++++++++++ hw/arm/virt-acpi-build.c | 8 ++++++++ 5 files changed, 49 insertions(+) diff --git a/include/hw/acpi/generic_event_device.h b/include/hw/acpi/generic_event_device.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/acpi/generic_event_device.h +++ b/include/hw/acpi/generic_event_device.h @@ -XXX,XX +XXX,XX @@ #include "hw/sysbus.h" #include "hw/acpi/memory_hotplug.h" +#include "hw/acpi/ghes.h" #define ACPI_POWER_BUTTON_DEVICE "PWRB" @@ -XXX,XX +XXX,XX @@ typedef struct AcpiGedState { GEDState ged_state; uint32_t ged_event_bitmap; qemu_irq irq; + AcpiGhesState ghes_state; } AcpiGedState; void build_ged_aml(Aml *table, const char* name, HotplugHandler *hotplug_dev, diff --git a/include/hw/acpi/ghes.h b/include/hw/acpi/ghes.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/acpi/ghes.h +++ b/include/hw/acpi/ghes.h @@ -XXX,XX +XXX,XX @@ enum { ACPI_HEST_SRC_ID_RESERVED, }; +typedef struct AcpiGhesState { + uint64_t ghes_addr_le; +} AcpiGhesState; + void build_ghes_error_table(GArray *hardware_errors, BIOSLinker *linker); void acpi_build_hest(GArray *table_data, BIOSLinker *linker); +void acpi_ghes_add_fw_cfg(AcpiGhesState *vms, FWCfgState *s, + GArray *hardware_errors); #endif diff --git a/hw/acpi/generic_event_device.c b/hw/acpi/generic_event_device.c index XXXXXXX..XXXXXXX 100644 --- a/hw/acpi/generic_event_device.c +++ b/hw/acpi/generic_event_device.c @@ -XXX,XX +XXX,XX @@ static const VMStateDescription vmstate_ged_state = { } }; +static bool ghes_needed(void *opaque) +{ + AcpiGedState *s = opaque; + return s->ghes_state.ghes_addr_le; +} + +static const VMStateDescription vmstate_ghes_state = { + .name = "acpi-ged/ghes", + .version_id = 1, + .minimum_version_id = 1, + .needed = ghes_needed, + .fields = (VMStateField[]) { + VMSTATE_STRUCT(ghes_state, AcpiGedState, 1, + vmstate_ghes_state, AcpiGhesState), + VMSTATE_END_OF_LIST() + } +}; + static const VMStateDescription vmstate_acpi_ged = { .name = "acpi-ged", .version_id = 1, @@ -XXX,XX +XXX,XX @@ static const VMStateDescription vmstate_acpi_ged = { }, .subsections = (const VMStateDescription * []) { &vmstate_memhp_state, + &vmstate_ghes_state, NULL } }; diff --git a/hw/acpi/ghes.c b/hw/acpi/ghes.c index XXXXXXX..XXXXXXX 100644 --- a/hw/acpi/ghes.c +++ b/hw/acpi/ghes.c @@ -XXX,XX +XXX,XX @@ #include "hw/acpi/ghes.h" #include "hw/acpi/aml-build.h" #include "qemu/error-report.h" +#include "hw/acpi/generic_event_device.h" +#include "hw/nvram/fw_cfg.h" #define ACPI_GHES_ERRORS_FW_CFG_FILE "etc/hardware_errors" #define ACPI_GHES_DATA_ADDR_FW_CFG_FILE "etc/hardware_errors_addr" @@ -XXX,XX +XXX,XX @@ void acpi_build_hest(GArray *table_data, BIOSLinker *linker) build_header(linker, table_data, (void *)(table_data->data + hest_start), "HEST", table_data->len - hest_start, 1, NULL, NULL); } + +void acpi_ghes_add_fw_cfg(AcpiGhesState *ags, FWCfgState *s, + GArray *hardware_error) +{ + /* Create a read-only fw_cfg file for GHES */ + fw_cfg_add_file(s, ACPI_GHES_ERRORS_FW_CFG_FILE, hardware_error->data, + hardware_error->len); + + /* Create a read-write fw_cfg file for Address */ + fw_cfg_add_file_callback(s, ACPI_GHES_DATA_ADDR_FW_CFG_FILE, NULL, NULL, + NULL, &(ags->ghes_addr_le), sizeof(ags->ghes_addr_le), false); +} diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/virt-acpi-build.c +++ b/hw/arm/virt-acpi-build.c @@ -XXX,XX +XXX,XX @@ void virt_acpi_setup(VirtMachineState *vms) { AcpiBuildTables tables; AcpiBuildState *build_state; + AcpiGedState *acpi_ged_state; if (!vms->fw_cfg) { trace_virt_acpi_setup(); @@ -XXX,XX +XXX,XX @@ void virt_acpi_setup(VirtMachineState *vms) fw_cfg_add_file(vms->fw_cfg, ACPI_BUILD_TPMLOG_FILE, tables.tcpalog->data, acpi_data_len(tables.tcpalog)); + if (vms->ras) { + assert(vms->acpi_dev); + acpi_ged_state = ACPI_GED(vms->acpi_dev); + acpi_ghes_add_fw_cfg(&acpi_ged_state->ghes_state, + vms->fw_cfg, tables.hardware_errors); + } + build_state->rsdp_mr = acpi_add_rom_blob(virt_acpi_build_update, build_state, tables.rsdp, ACPI_BUILD_RSDP_FILE, 0); -- 2.20.1 From: Dongjiu Geng <gengdongjiu@huawei.com> kvm_hwpoison_page_add() and kvm_unpoison_all() will both be used by X86 and ARM platforms, so moving them into "accel/kvm/kvm-all.c" to avoid duplicate code. For architectures that don't use the poison-list functionality the reset handler will harmlessly do nothing, so let's register the kvm_unpoison_all() function in the generic kvm_init() function. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Dongjiu Geng <gengdongjiu@huawei.com> Signed-off-by: Xiang Zheng <zhengxiang9@huawei.com> Acked-by: Xiang Zheng <zhengxiang9@huawei.com> Message-id: 20200512030609.19593-8-gengdongjiu@huawei.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/sysemu/kvm_int.h | 12 ++++++++++++ accel/kvm/kvm-all.c | 36 ++++++++++++++++++++++++++++++++++++ target/i386/kvm.c | 36 ------------------------------------ 3 files changed, 48 insertions(+), 36 deletions(-) diff --git a/include/sysemu/kvm_int.h b/include/sysemu/kvm_int.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/kvm_int.h +++ b/include/sysemu/kvm_int.h @@ -XXX,XX +XXX,XX @@ void kvm_memory_listener_register(KVMState *s, KVMMemoryListener *kml, AddressSpace *as, int as_id); void kvm_set_max_memslot_size(hwaddr max_slot_size); + +/** + * kvm_hwpoison_page_add: + * + * Parameters: + * @ram_addr: the address in the RAM for the poisoned page + * + * Add a poisoned page to the list + * + * Return: None. + */ +void kvm_hwpoison_page_add(ram_addr_t ram_addr); #endif diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c index XXXXXXX..XXXXXXX 100644 --- a/accel/kvm/kvm-all.c +++ b/accel/kvm/kvm-all.c @@ -XXX,XX +XXX,XX @@ #include "qapi/visitor.h" #include "qapi/qapi-types-common.h" #include "qapi/qapi-visit-common.h" +#include "sysemu/reset.h" #include "hw/boards.h" @@ -XXX,XX +XXX,XX @@ int kvm_vm_check_extension(KVMState *s, unsigned int extension) return ret; } +typedef struct HWPoisonPage { + ram_addr_t ram_addr; + QLIST_ENTRY(HWPoisonPage) list; +} HWPoisonPage; + +static QLIST_HEAD(, HWPoisonPage) hwpoison_page_list = + QLIST_HEAD_INITIALIZER(hwpoison_page_list); + +static void kvm_unpoison_all(void *param) +{ + HWPoisonPage *page, *next_page; + + QLIST_FOREACH_SAFE(page, &hwpoison_page_list, list, next_page) { + QLIST_REMOVE(page, list); + qemu_ram_remap(page->ram_addr, TARGET_PAGE_SIZE); + g_free(page); + } +} + +void kvm_hwpoison_page_add(ram_addr_t ram_addr) +{ + HWPoisonPage *page; + + QLIST_FOREACH(page, &hwpoison_page_list, list) { + if (page->ram_addr == ram_addr) { + return; + } + } + page = g_new(HWPoisonPage, 1); + page->ram_addr = ram_addr; + QLIST_INSERT_HEAD(&hwpoison_page_list, page, list); +} + static uint32_t adjust_ioeventfd_endianness(uint32_t val, uint32_t size) { #if defined(HOST_WORDS_BIGENDIAN) != defined(TARGET_WORDS_BIGENDIAN) @@ -XXX,XX +XXX,XX @@ static int kvm_init(MachineState *ms) s->kernel_irqchip_split = mc->default_kernel_irqchip_split ? ON_OFF_AUTO_ON : ON_OFF_AUTO_OFF; } + qemu_register_reset(kvm_unpoison_all, NULL); + if (s->kernel_irqchip_allowed) { kvm_irqchip_create(s); } diff --git a/target/i386/kvm.c b/target/i386/kvm.c index XXXXXXX..XXXXXXX 100644 --- a/target/i386/kvm.c +++ b/target/i386/kvm.c @@ -XXX,XX +XXX,XX @@ #include "sysemu/sysemu.h" #include "sysemu/hw_accel.h" #include "sysemu/kvm_int.h" -#include "sysemu/reset.h" #include "sysemu/runstate.h" #include "kvm_i386.h" #include "hyperv.h" @@ -XXX,XX +XXX,XX @@ uint64_t kvm_arch_get_supported_msr_feature(KVMState *s, uint32_t index) } } - -typedef struct HWPoisonPage { - ram_addr_t ram_addr; - QLIST_ENTRY(HWPoisonPage) list; -} HWPoisonPage; - -static QLIST_HEAD(, HWPoisonPage) hwpoison_page_list = - QLIST_HEAD_INITIALIZER(hwpoison_page_list); - -static void kvm_unpoison_all(void *param) -{ - HWPoisonPage *page, *next_page; - - QLIST_FOREACH_SAFE(page, &hwpoison_page_list, list, next_page) { - QLIST_REMOVE(page, list); - qemu_ram_remap(page->ram_addr, TARGET_PAGE_SIZE); - g_free(page); - } -} - -static void kvm_hwpoison_page_add(ram_addr_t ram_addr) -{ - HWPoisonPage *page; - - QLIST_FOREACH(page, &hwpoison_page_list, list) { - if (page->ram_addr == ram_addr) { - return; - } - } - page = g_new(HWPoisonPage, 1); - page->ram_addr = ram_addr; - QLIST_INSERT_HEAD(&hwpoison_page_list, page, list); -} - static int kvm_get_mce_cap_supported(KVMState *s, uint64_t *mce_cap, int *max_banks) { @@ -XXX,XX +XXX,XX @@ int kvm_arch_init(MachineState *ms, KVMState *s) fprintf(stderr, "e820_add_entry() table is full\n"); return ret; } - qemu_register_reset(kvm_unpoison_all, NULL); shadow_mem = object_property_get_int(OBJECT(s), "kvm-shadow-mem", &error_abort); if (shadow_mem != -1) { -- 2.20.1 From: Dongjiu Geng <gengdongjiu@huawei.com> kvm_arch_on_sigbus_vcpu() error injection uses source_id as index in etc/hardware_errors to find out Error Status Data Block entry corresponding to error source. So supported source_id values should be assigned here and not be changed afterwards to make sure that guest will write error into expected Error Status Data Block. Before QEMU writes a new error to ACPI table, it will check whether previous error has been acknowledged. If not acknowledged, the new errors will be ignored and not be recorded. For the errors section type, QEMU simulate it to memory section error. Signed-off-by: Dongjiu Geng <gengdongjiu@huawei.com> Signed-off-by: Xiang Zheng <zhengxiang9@huawei.com> Reviewed-by: Igor Mammedov <imammedo@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Message-id: 20200512030609.19593-9-gengdongjiu@huawei.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/acpi/ghes.h | 1 + hw/acpi/ghes.c | 219 +++++++++++++++++++++++++++++++++++++++++ 2 files changed, 220 insertions(+) diff --git a/include/hw/acpi/ghes.h b/include/hw/acpi/ghes.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/acpi/ghes.h +++ b/include/hw/acpi/ghes.h @@ -XXX,XX +XXX,XX @@ void build_ghes_error_table(GArray *hardware_errors, BIOSLinker *linker); void acpi_build_hest(GArray *table_data, BIOSLinker *linker); void acpi_ghes_add_fw_cfg(AcpiGhesState *vms, FWCfgState *s, GArray *hardware_errors); +int acpi_ghes_record_errors(uint8_t notify, uint64_t error_physical_addr); #endif diff --git a/hw/acpi/ghes.c b/hw/acpi/ghes.c index XXXXXXX..XXXXXXX 100644 --- a/hw/acpi/ghes.c +++ b/hw/acpi/ghes.c @@ -XXX,XX +XXX,XX @@ #include "qemu/error-report.h" #include "hw/acpi/generic_event_device.h" #include "hw/nvram/fw_cfg.h" +#include "qemu/uuid.h" #define ACPI_GHES_ERRORS_FW_CFG_FILE "etc/hardware_errors" #define ACPI_GHES_DATA_ADDR_FW_CFG_FILE "etc/hardware_errors_addr" @@ -XXX,XX +XXX,XX @@ /* Address offset in Generic Address Structure(GAS) */ #define GAS_ADDR_OFFSET 4 +/* + * The total size of Generic Error Data Entry + * ACPI 6.1/6.2: 18.3.2.7.1 Generic Error Data, + * Table 18-343 Generic Error Data Entry + */ +#define ACPI_GHES_DATA_LENGTH 72 + +/* The memory section CPER size, UEFI 2.6: N.2.5 Memory Error Section */ +#define ACPI_GHES_MEM_CPER_LENGTH 80 + +/* Masks for block_status flags */ +#define ACPI_GEBS_UNCORRECTABLE 1 + +/* + * Total size for Generic Error Status Block except Generic Error Data Entries + * ACPI 6.2: 18.3.2.7.1 Generic Error Data, + * Table 18-380 Generic Error Status Block + */ +#define ACPI_GHES_GESB_SIZE 20 + +/* + * Values for error_severity field + */ +enum AcpiGenericErrorSeverity { + ACPI_CPER_SEV_RECOVERABLE = 0, + ACPI_CPER_SEV_FATAL = 1, + ACPI_CPER_SEV_CORRECTED = 2, + ACPI_CPER_SEV_NONE = 3, +}; + /* * Hardware Error Notification * ACPI 4.0: 17.3.2.7 Hardware Error Notification @@ -XXX,XX +XXX,XX @@ static void build_ghes_hw_error_notification(GArray *table, const uint8_t type) build_append_int_noprefix(table, 0, 4); } +/* + * Generic Error Data Entry + * ACPI 6.1: 18.3.2.7.1 Generic Error Data + */ +static void acpi_ghes_generic_error_data(GArray *table, + const uint8_t *section_type, uint32_t error_severity, + uint8_t validation_bits, uint8_t flags, + uint32_t error_data_length, QemuUUID fru_id, + uint64_t time_stamp) +{ + const uint8_t fru_text[20] = {0}; + + /* Section Type */ + g_array_append_vals(table, section_type, 16); + + /* Error Severity */ + build_append_int_noprefix(table, error_severity, 4); + /* Revision */ + build_append_int_noprefix(table, 0x300, 2); + /* Validation Bits */ + build_append_int_noprefix(table, validation_bits, 1); + /* Flags */ + build_append_int_noprefix(table, flags, 1); + /* Error Data Length */ + build_append_int_noprefix(table, error_data_length, 4); + + /* FRU Id */ + g_array_append_vals(table, fru_id.data, ARRAY_SIZE(fru_id.data)); + + /* FRU Text */ + g_array_append_vals(table, fru_text, sizeof(fru_text)); + + /* Timestamp */ + build_append_int_noprefix(table, time_stamp, 8); +} + +/* + * Generic Error Status Block + * ACPI 6.1: 18.3.2.7.1 Generic Error Data + */ +static void acpi_ghes_generic_error_status(GArray *table, uint32_t block_status, + uint32_t raw_data_offset, uint32_t raw_data_length, + uint32_t data_length, uint32_t error_severity) +{ + /* Block Status */ + build_append_int_noprefix(table, block_status, 4); + /* Raw Data Offset */ + build_append_int_noprefix(table, raw_data_offset, 4); + /* Raw Data Length */ + build_append_int_noprefix(table, raw_data_length, 4); + /* Data Length */ + build_append_int_noprefix(table, data_length, 4); + /* Error Severity */ + build_append_int_noprefix(table, error_severity, 4); +} + +/* UEFI 2.6: N.2.5 Memory Error Section */ +static void acpi_ghes_build_append_mem_cper(GArray *table, + uint64_t error_physical_addr) +{ + /* + * Memory Error Record + */ + + /* Validation Bits */ + build_append_int_noprefix(table, + (1ULL << 14) | /* Type Valid */ + (1ULL << 1) /* Physical Address Valid */, + 8); + /* Error Status */ + build_append_int_noprefix(table, 0, 8); + /* Physical Address */ + build_append_int_noprefix(table, error_physical_addr, 8); + /* Skip all the detailed information normally found in such a record */ + build_append_int_noprefix(table, 0, 48); + /* Memory Error Type */ + build_append_int_noprefix(table, 0 /* Unknown error */, 1); + /* Skip all the detailed information normally found in such a record */ + build_append_int_noprefix(table, 0, 7); +} + +static int acpi_ghes_record_mem_error(uint64_t error_block_address, + uint64_t error_physical_addr) +{ + GArray *block; + + /* Memory Error Section Type */ + const uint8_t uefi_cper_mem_sec[] = + UUID_LE(0xA5BC1114, 0x6F64, 0x4EDE, 0xB8, 0x63, 0x3E, 0x83, \ + 0xED, 0x7C, 0x83, 0xB1); + + /* invalid fru id: ACPI 4.0: 17.3.2.6.1 Generic Error Data, + * Table 17-13 Generic Error Data Entry + */ + QemuUUID fru_id = {}; + uint32_t data_length; + + block = g_array_new(false, true /* clear */, 1); + + /* This is the length if adding a new generic error data entry*/ + data_length = ACPI_GHES_DATA_LENGTH + ACPI_GHES_MEM_CPER_LENGTH; + + /* + * Check whether it will run out of the preallocated memory if adding a new + * generic error data entry + */ + if ((data_length + ACPI_GHES_GESB_SIZE) > ACPI_GHES_MAX_RAW_DATA_LENGTH) { + error_report("Not enough memory to record new CPER!!!"); + g_array_free(block, true); + return -1; + } + + /* Build the new generic error status block header */ + acpi_ghes_generic_error_status(block, ACPI_GEBS_UNCORRECTABLE, + 0, 0, data_length, ACPI_CPER_SEV_RECOVERABLE); + + /* Build this new generic error data entry header */ + acpi_ghes_generic_error_data(block, uefi_cper_mem_sec, + ACPI_CPER_SEV_RECOVERABLE, 0, 0, + ACPI_GHES_MEM_CPER_LENGTH, fru_id, 0); + + /* Build the memory section CPER for above new generic error data entry */ + acpi_ghes_build_append_mem_cper(block, error_physical_addr); + + /* Write the generic error data entry into guest memory */ + cpu_physical_memory_write(error_block_address, block->data, block->len); + + g_array_free(block, true); + + return 0; +} + /* * Build table for the hardware error fw_cfg blob. * Initialize "etc/hardware_errors" and "etc/hardware_errors_addr" fw_cfg blobs. @@ -XXX,XX +XXX,XX @@ void acpi_ghes_add_fw_cfg(AcpiGhesState *ags, FWCfgState *s, fw_cfg_add_file_callback(s, ACPI_GHES_DATA_ADDR_FW_CFG_FILE, NULL, NULL, NULL, &(ags->ghes_addr_le), sizeof(ags->ghes_addr_le), false); } + +int acpi_ghes_record_errors(uint8_t source_id, uint64_t physical_address) +{ + uint64_t error_block_addr, read_ack_register_addr, read_ack_register = 0; + uint64_t start_addr; + bool ret = -1; + AcpiGedState *acpi_ged_state; + AcpiGhesState *ags; + + assert(source_id < ACPI_HEST_SRC_ID_RESERVED); + + acpi_ged_state = ACPI_GED(object_resolve_path_type("", TYPE_ACPI_GED, + NULL)); + g_assert(acpi_ged_state); + ags = &acpi_ged_state->ghes_state; + + start_addr = le64_to_cpu(ags->ghes_addr_le); + + if (physical_address) { + + if (source_id < ACPI_HEST_SRC_ID_RESERVED) { + start_addr += source_id * sizeof(uint64_t); + } + + cpu_physical_memory_read(start_addr, &error_block_addr, + sizeof(error_block_addr)); + + error_block_addr = le64_to_cpu(error_block_addr); + + read_ack_register_addr = start_addr + + ACPI_GHES_ERROR_SOURCE_COUNT * sizeof(uint64_t); + + cpu_physical_memory_read(read_ack_register_addr, + &read_ack_register, sizeof(read_ack_register)); + + /* zero means OSPM does not acknowledge the error */ + if (!read_ack_register) { + error_report("OSPM does not acknowledge previous error," + " so can not record CPER for current error anymore"); + } else if (error_block_addr) { + read_ack_register = cpu_to_le64(0); + /* + * Clear the Read Ack Register, OSPM will write it to 1 when + * it acknowledges this error. + */ + cpu_physical_memory_write(read_ack_register_addr, + &read_ack_register, sizeof(uint64_t)); + + ret = acpi_ghes_record_mem_error(error_block_addr, + physical_address); + } else + error_report("can not find Generic Error Status Block"); + } + + return ret; +} -- 2.20.1 From: Dongjiu Geng <gengdongjiu@huawei.com> Add a SIGBUS signal handler. In this handler, it checks the SIGBUS type, translates the host VA delivered by host to guest PA, then fills this PA to guest APEI GHES memory, then notifies guest according to the SIGBUS type. When guest accesses the poisoned memory, it will generate a Synchronous External Abort(SEA). Then host kernel gets an APEI notification and calls memory_failure() to unmapped the affected page in stage 2, finally returns to guest. Guest continues to access the PG_hwpoison page, it will trap to KVM as stage2 fault, then a SIGBUS_MCEERR_AR synchronous signal is delivered to Qemu, Qemu records this error address into guest APEI GHES memory and notifes guest using Synchronous-External-Abort(SEA). In order to inject a vSEA, we introduce the kvm_inject_arm_sea() function in which we can setup the type of exception and the syndrome information. When switching to guest, the target vcpu will jump to the synchronous external abort vector table entry. The ESR_ELx.DFSC is set to synchronous external abort(0x10), and the ESR_ELx.FnV is set to not valid(0x1), which will tell guest that FAR is not valid and hold an UNKNOWN value. These values will be set to KVM register structures through KVM_SET_ONE_REG IOCTL. Signed-off-by: Dongjiu Geng <gengdongjiu@huawei.com> Signed-off-by: Xiang Zheng <zhengxiang9@huawei.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Acked-by: Xiang Zheng <zhengxiang9@huawei.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Igor Mammedov <imammedo@redhat.com> Message-id: 20200512030609.19593-10-gengdongjiu@huawei.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/sysemu/kvm.h | 3 +- target/arm/cpu.h | 4 +++ target/arm/internals.h | 5 +-- target/i386/cpu.h | 2 ++ target/arm/helper.c | 2 +- target/arm/kvm64.c | 77 +++++++++++++++++++++++++++++++++++++++++ target/arm/tlb_helper.c | 2 +- 7 files changed, 89 insertions(+), 6 deletions(-) diff --git a/include/sysemu/kvm.h b/include/sysemu/kvm.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/kvm.h +++ b/include/sysemu/kvm.h @@ -XXX,XX +XXX,XX @@ bool kvm_vcpu_id_is_valid(int vcpu_id); /* Returns VCPU ID to be used on KVM_CREATE_VCPU ioctl() */ unsigned long kvm_arch_vcpu_id(CPUState *cpu); -#ifdef TARGET_I386 -#define KVM_HAVE_MCE_INJECTION 1 +#ifdef KVM_HAVE_MCE_INJECTION void kvm_arch_on_sigbus_vcpu(CPUState *cpu, int code, void *addr); #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 @@ /* ARM processors have a weak memory model */ #define TCG_GUEST_DEFAULT_MO (0) +#ifdef TARGET_AARCH64 +#define KVM_HAVE_MCE_INJECTION 1 +#endif + #define EXCP_UDEF 1 /* undefined instruction */ #define EXCP_SWI 2 /* software interrupt */ #define EXCP_PREFETCH_ABORT 3 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 syn_insn_abort(int same_el, int ea, int s1ptw, int fsc) | ARM_EL_IL | (ea << 9) | (s1ptw << 7) | fsc; } -static inline uint32_t syn_data_abort_no_iss(int same_el, +static inline uint32_t syn_data_abort_no_iss(int same_el, int fnv, int ea, int cm, int s1ptw, int wnr, int fsc) { return (EC_DATAABORT << ARM_EL_EC_SHIFT) | (same_el << ARM_EL_EC_SHIFT) | ARM_EL_IL - | (ea << 9) | (cm << 8) | (s1ptw << 7) | (wnr << 6) | fsc; + | (fnv << 10) | (ea << 9) | (cm << 8) | (s1ptw << 7) + | (wnr << 6) | fsc; } static inline uint32_t syn_data_abort_with_iss(int same_el, diff --git a/target/i386/cpu.h b/target/i386/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/i386/cpu.h +++ b/target/i386/cpu.h @@ -XXX,XX +XXX,XX @@ /* The x86 has a strong memory model with some store-after-load re-ordering */ #define TCG_GUEST_DEFAULT_MO (TCG_MO_ALL & ~TCG_MO_ST_LD) +#define KVM_HAVE_MCE_INJECTION 1 + /* Maximum instruction code size */ #define TARGET_MAX_INSN_SIZE 16 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 uint64_t do_ats_write(CPUARMState *env, uint64_t value, * Report exception with ESR indicating a fault due to a * translation table walk for a cache maintenance instruction. */ - syn = syn_data_abort_no_iss(current_el == target_el, + syn = syn_data_abort_no_iss(current_el == target_el, 0, fi.ea, 1, fi.s1ptw, 1, fsc); env->exception.vaddress = value; env->exception.fsr = fsr; diff --git a/target/arm/kvm64.c b/target/arm/kvm64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/kvm64.c +++ b/target/arm/kvm64.c @@ -XXX,XX +XXX,XX @@ #include "sysemu/kvm_int.h" #include "kvm_arm.h" #include "internals.h" +#include "hw/acpi/acpi.h" +#include "hw/acpi/ghes.h" +#include "hw/arm/virt.h" static bool have_guest_debug; @@ -XXX,XX +XXX,XX @@ int kvm_arm_cpreg_level(uint64_t regidx) return KVM_PUT_RUNTIME_STATE; } +/* Callers must hold the iothread mutex lock */ +static void kvm_inject_arm_sea(CPUState *c) +{ + ARMCPU *cpu = ARM_CPU(c); + CPUARMState *env = &cpu->env; + CPUClass *cc = CPU_GET_CLASS(c); + uint32_t esr; + bool same_el; + + c->exception_index = EXCP_DATA_ABORT; + env->exception.target_el = 1; + + /* + * Set the DFSC to synchronous external abort and set FnV to not valid, + * this will tell guest the FAR_ELx is UNKNOWN for this abort. + */ + same_el = arm_current_el(env) == env->exception.target_el; + esr = syn_data_abort_no_iss(same_el, 1, 0, 0, 0, 0, 0x10); + + env->exception.syndrome = esr; + + cc->do_interrupt(c); +} + #define AARCH64_CORE_REG(x) (KVM_REG_ARM64 | KVM_REG_SIZE_U64 | \ KVM_REG_ARM_CORE | KVM_REG_ARM_CORE_REG(x)) @@ -XXX,XX +XXX,XX @@ int kvm_arch_get_registers(CPUState *cs) return ret; } +void kvm_arch_on_sigbus_vcpu(CPUState *c, int code, void *addr) +{ + ram_addr_t ram_addr; + hwaddr paddr; + Object *obj = qdev_get_machine(); + VirtMachineState *vms = VIRT_MACHINE(obj); + bool acpi_enabled = virt_is_acpi_enabled(vms); + + assert(code == BUS_MCEERR_AR || code == BUS_MCEERR_AO); + + if (acpi_enabled && addr && + object_property_get_bool(obj, "ras", NULL)) { + ram_addr = qemu_ram_addr_from_host(addr); + if (ram_addr != RAM_ADDR_INVALID && + kvm_physical_memory_addr_from_host(c->kvm_state, addr, &paddr)) { + kvm_hwpoison_page_add(ram_addr); + /* + * If this is a BUS_MCEERR_AR, we know we have been called + * synchronously from the vCPU thread, so we can easily + * synchronize the state and inject an error. + * + * TODO: we currently don't tell the guest at all about + * BUS_MCEERR_AO. In that case we might either be being + * called synchronously from the vCPU thread, or a bit + * later from the main thread, so doing the injection of + * the error would be more complicated. + */ + if (code == BUS_MCEERR_AR) { + kvm_cpu_synchronize_state(c); + if (!acpi_ghes_record_errors(ACPI_HEST_SRC_ID_SEA, paddr)) { + kvm_inject_arm_sea(c); + } else { + error_report("failed to record the error"); + abort(); + } + } + return; + } + if (code == BUS_MCEERR_AO) { + error_report("Hardware memory error at addr %p for memory used by " + "QEMU itself instead of guest system!", addr); + } + } + + if (code == BUS_MCEERR_AR) { + error_report("Hardware memory error!"); + exit(1); + } +} + /* C6.6.29 BRK instruction */ static const uint32_t brk_insn = 0xd4200000; 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 @@ static inline uint32_t merge_syn_data_abort(uint32_t template_syn, * ISV field. */ if (!(template_syn & ARM_EL_ISV) || target_el != 2 || s1ptw) { - syn = syn_data_abort_no_iss(same_el, + syn = syn_data_abort_no_iss(same_el, 0, ea, 0, s1ptw, is_write, fsc); } else { /* -- 2.20.1 From: Dongjiu Geng <gengdongjiu@huawei.com> I and Xiang are willing to review the APEI-related patches and volunteer as the reviewers for the HEST/GHES part. Signed-off-by: Dongjiu Geng <gengdongjiu@huawei.com> Signed-off-by: Xiang Zheng <zhengxiang9@huawei.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Acked-by: Michael S. Tsirkin <mst@redhat.com> Message-id: 20200512030609.19593-11-gengdongjiu@huawei.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- MAINTAINERS | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ F: tests/qtest/bios-tables-test.c F: tests/qtest/acpi-utils.[hc] F: tests/data/acpi/ +ACPI/HEST/GHES +R: Dongjiu Geng <gengdongjiu@huawei.com> +R: Xiang Zheng <zhengxiang9@huawei.com> +L: qemu-arm@nongnu.org +S: Maintained +F: hw/acpi/ghes.c +F: include/hw/acpi/ghes.h +F: docs/specs/acpi_hest_ghes.rst + ppc4xx M: David Gibson <david@gibson.dropbear.id.au> L: qemu-ppc@nongnu.org -- 2.20.1 Convert the Neon VQRDMLAH and VQRDMLSH insns in the 3-reg-same group to decodetree. These don't use do_3same() because they want to operate on VFP double registers, whose offsets are different from the neon_reg_offset() calculations do_3same does. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-2-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 3 +++ target/arm/translate-neon.inc.c | 15 +++++++++++++++ target/arm/translate.c | 14 ++------------ 3 files changed, 20 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 @@ 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 + +VQRDMLAH_3s 1111 001 1 0 . .. .... .... 1011 ... 1 .... @3same +VQRDMLSH_3s 1111 001 1 0 . .. .... .... 1100 ... 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 bool trans_VMUL_p_3s(DisasContext *s, arg_3same *a) } return do_3same(s, a, gen_VMUL_p_3s); } + +#define DO_VQRDMLAH(INSN, FUNC) \ + static bool trans_##INSN##_3s(DisasContext *s, arg_3same *a) \ + { \ + if (!dc_isar_feature(aa32_rdm, s)) { \ + return false; \ + } \ + if (a->size != 1 && a->size != 2) { \ + return false; \ + } \ + return do_3same(s, a, FUNC); \ + } + +DO_VQRDMLAH(VQRDMLAH, gen_gvec_sqrdmlah_qc) +DO_VQRDMLAH(VQRDMLSH, gen_gvec_sqrdmlsh_qc) 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) if (!u) { break; /* VPADD */ } - /* VQRDMLAH */ - if (dc_isar_feature(aa32_rdm, s) && (size == 1 || size == 2)) { - gen_gvec_sqrdmlah_qc(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - return 0; - } + /* VQRDMLAH : handled by decodetree */ return 1; case NEON_3R_VFM_VQRDMLSH: @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) } break; } - /* VQRDMLSH */ - if (dc_isar_feature(aa32_rdm, s) && (size == 1 || size == 2)) { - gen_gvec_sqrdmlsh_qc(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - return 0; - } + /* VQRDMLSH : handled by decodetree */ return 1; case NEON_3R_VABD: -- 2.20.1 Convert the Neon SHA instructions in the 3-reg-same group to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-3-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 10 +++ target/arm/translate-neon.inc.c | 139 ++++++++++++++++++++++++++++++++ target/arm/translate.c | 46 +---------- 3 files changed, 151 insertions(+), 44 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 @@ VMUL_3s 1111 001 0 0 . .. .... .... 1001 . . . 1 .... @3same VMUL_p_3s 1111 001 1 0 . .. .... .... 1001 . . . 1 .... @3same VQRDMLAH_3s 1111 001 1 0 . .. .... .... 1011 ... 1 .... @3same + +SHA1_3s 1111 001 0 0 . optype:2 .... .... 1100 . 1 . 0 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp +SHA256H_3s 1111 001 1 0 . 00 .... .... 1100 . 1 . 0 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp +SHA256H2_3s 1111 001 1 0 . 01 .... .... 1100 . 1 . 0 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp +SHA256SU1_3s 1111 001 1 0 . 10 .... .... 1100 . 1 . 0 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp + VQRDMLSH_3s 1111 001 1 0 . .. .... .... 1100 ... 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 bool trans_VMUL_p_3s(DisasContext *s, arg_3same *a) DO_VQRDMLAH(VQRDMLAH, gen_gvec_sqrdmlah_qc) DO_VQRDMLAH(VQRDMLSH, gen_gvec_sqrdmlsh_qc) + +static bool trans_SHA1_3s(DisasContext *s, arg_SHA1_3s *a) +{ + TCGv_ptr ptr1, ptr2, ptr3; + TCGv_i32 tmp; + + if (!arm_dc_feature(s, ARM_FEATURE_NEON) || + !dc_isar_feature(aa32_sha1, 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) & 1) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + ptr1 = vfp_reg_ptr(true, a->vd); + ptr2 = vfp_reg_ptr(true, a->vn); + ptr3 = vfp_reg_ptr(true, a->vm); + tmp = tcg_const_i32(a->optype); + gen_helper_crypto_sha1_3reg(ptr1, ptr2, ptr3, tmp); + tcg_temp_free_i32(tmp); + tcg_temp_free_ptr(ptr1); + tcg_temp_free_ptr(ptr2); + tcg_temp_free_ptr(ptr3); + + return true; +} + +static bool trans_SHA256H_3s(DisasContext *s, arg_SHA256H_3s *a) +{ + TCGv_ptr ptr1, ptr2, ptr3; + + if (!arm_dc_feature(s, ARM_FEATURE_NEON) || + !dc_isar_feature(aa32_sha2, 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) & 1) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + ptr1 = vfp_reg_ptr(true, a->vd); + ptr2 = vfp_reg_ptr(true, a->vn); + ptr3 = vfp_reg_ptr(true, a->vm); + gen_helper_crypto_sha256h(ptr1, ptr2, ptr3); + tcg_temp_free_ptr(ptr1); + tcg_temp_free_ptr(ptr2); + tcg_temp_free_ptr(ptr3); + + return true; +} + +static bool trans_SHA256H2_3s(DisasContext *s, arg_SHA256H2_3s *a) +{ + TCGv_ptr ptr1, ptr2, ptr3; + + if (!arm_dc_feature(s, ARM_FEATURE_NEON) || + !dc_isar_feature(aa32_sha2, 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) & 1) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + ptr1 = vfp_reg_ptr(true, a->vd); + ptr2 = vfp_reg_ptr(true, a->vn); + ptr3 = vfp_reg_ptr(true, a->vm); + gen_helper_crypto_sha256h2(ptr1, ptr2, ptr3); + tcg_temp_free_ptr(ptr1); + tcg_temp_free_ptr(ptr2); + tcg_temp_free_ptr(ptr3); + + return true; +} + +static bool trans_SHA256SU1_3s(DisasContext *s, arg_SHA256SU1_3s *a) +{ + TCGv_ptr ptr1, ptr2, ptr3; + + if (!arm_dc_feature(s, ARM_FEATURE_NEON) || + !dc_isar_feature(aa32_sha2, 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) & 1) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + ptr1 = vfp_reg_ptr(true, a->vd); + ptr2 = vfp_reg_ptr(true, a->vn); + ptr3 = vfp_reg_ptr(true, a->vm); + gen_helper_crypto_sha256su1(ptr1, ptr2, ptr3); + tcg_temp_free_ptr(ptr1); + tcg_temp_free_ptr(ptr2); + tcg_temp_free_ptr(ptr3); + + 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) int vec_size; uint32_t imm; TCGv_i32 tmp, tmp2, tmp3, tmp4, tmp5; - TCGv_ptr ptr1, ptr2, ptr3; + TCGv_ptr ptr1, ptr2; TCGv_i64 tmp64; if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) return 1; } switch (op) { - case NEON_3R_SHA: - /* The SHA-1/SHA-256 3-register instructions require special - * treatment here, as their size field is overloaded as an - * op type selector, and they all consume their input in a - * single pass. - */ - if (!q) { - return 1; - } - if (!u) { /* SHA-1 */ - if (!dc_isar_feature(aa32_sha1, s)) { - return 1; - } - ptr1 = vfp_reg_ptr(true, rd); - ptr2 = vfp_reg_ptr(true, rn); - ptr3 = vfp_reg_ptr(true, rm); - tmp4 = tcg_const_i32(size); - gen_helper_crypto_sha1_3reg(ptr1, ptr2, ptr3, tmp4); - tcg_temp_free_i32(tmp4); - } else { /* SHA-256 */ - if (!dc_isar_feature(aa32_sha2, s) || size == 3) { - return 1; - } - ptr1 = vfp_reg_ptr(true, rd); - ptr2 = vfp_reg_ptr(true, rn); - ptr3 = vfp_reg_ptr(true, rm); - switch (size) { - case 0: - gen_helper_crypto_sha256h(ptr1, ptr2, ptr3); - break; - case 1: - gen_helper_crypto_sha256h2(ptr1, ptr2, ptr3); - break; - case 2: - gen_helper_crypto_sha256su1(ptr1, ptr2, ptr3); - break; - } - } - tcg_temp_free_ptr(ptr1); - tcg_temp_free_ptr(ptr2); - tcg_temp_free_ptr(ptr3); - return 0; - case NEON_3R_VPADD_VQRDMLAH: if (!u) { break; /* VPADD */ @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) case NEON_3R_VMUL: case NEON_3R_VML: case NEON_3R_VSHL: + case NEON_3R_SHA: /* Already handled by decodetree */ return 1; } -- 2.20.1 Convert the 64-bit element insns in the 3-reg-same group to decodetree. This covers VQSHL, VRSHL and VQRSHL where size==0b11. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-4-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 13 +++++++++++ target/arm/translate-neon.inc.c | 24 +++++++++++++++++++++ target/arm/translate.c | 38 ++------------------------------- 3 files changed, 39 insertions(+), 36 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 @@ VCGE_U_3s 1111 001 1 0 . .. .... .... 0011 . . . 1 .... @3same VSHL_S_3s 1111 001 0 0 . .. .... .... 0100 . . . 0 .... @3same_rev VSHL_U_3s 1111 001 1 0 . .. .... .... 0100 . . . 0 .... @3same_rev +# Insns operating on 64-bit elements (size!=0b11 handled elsewhere) +# The _rev suffix indicates that Vn and Vm are reversed (as explained +# by the comment for the @3same_rev format). +@3same_64_rev .... ... . . . 11 .... .... .... . q:1 . . .... \ + &3same vm=%vn_dp vn=%vm_dp vd=%vd_dp size=3 + +VQSHL_S64_3s 1111 001 0 0 . .. .... .... 0100 . . . 1 .... @3same_64_rev +VQSHL_U64_3s 1111 001 1 0 . .. .... .... 0100 . . . 1 .... @3same_64_rev +VRSHL_S64_3s 1111 001 0 0 . .. .... .... 0101 . . . 0 .... @3same_64_rev +VRSHL_U64_3s 1111 001 1 0 . .. .... .... 0101 . . . 0 .... @3same_64_rev +VQRSHL_S64_3s 1111 001 0 0 . .. .... .... 0101 . . . 1 .... @3same_64_rev +VQRSHL_U64_3s 1111 001 1 0 . .. .... .... 0101 . . . 1 .... @3same_64_rev + 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 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_SHA256SU1_3s(DisasContext *s, arg_SHA256SU1_3s *a) return true; } + +#define DO_3SAME_64(INSN, FUNC) \ + 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) \ + { \ + static const GVecGen3 op = { .fni8 = FUNC }; \ + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, oprsz, maxsz, &op); \ + } \ + DO_3SAME(INSN, gen_##INSN##_3s) + +#define DO_3SAME_64_ENV(INSN, FUNC) \ + static void gen_##INSN##_elt(TCGv_i64 d, TCGv_i64 n, TCGv_i64 m) \ + { \ + FUNC(d, cpu_env, n, m); \ + } \ + DO_3SAME_64(INSN, gen_##INSN##_elt) + +DO_3SAME_64(VRSHL_S64, gen_helper_neon_rshl_s64) +DO_3SAME_64(VRSHL_U64, gen_helper_neon_rshl_u64) +DO_3SAME_64_ENV(VQSHL_S64, gen_helper_neon_qshl_s64) +DO_3SAME_64_ENV(VQSHL_U64, gen_helper_neon_qshl_u64) +DO_3SAME_64_ENV(VQRSHL_S64, gen_helper_neon_qrshl_s64) +DO_3SAME_64_ENV(VQRSHL_U64, gen_helper_neon_qrshl_u64) 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) } if (size == 3) { - /* 64-bit element instructions. */ - for (pass = 0; pass < (q ? 2 : 1); pass++) { - neon_load_reg64(cpu_V0, rn + pass); - neon_load_reg64(cpu_V1, rm + pass); - switch (op) { - case NEON_3R_VQSHL: - if (u) { - gen_helper_neon_qshl_u64(cpu_V0, cpu_env, - cpu_V1, cpu_V0); - } else { - gen_helper_neon_qshl_s64(cpu_V0, cpu_env, - cpu_V1, cpu_V0); - } - break; - case NEON_3R_VRSHL: - if (u) { - gen_helper_neon_rshl_u64(cpu_V0, cpu_V1, cpu_V0); - } else { - gen_helper_neon_rshl_s64(cpu_V0, cpu_V1, cpu_V0); - } - break; - case NEON_3R_VQRSHL: - if (u) { - gen_helper_neon_qrshl_u64(cpu_V0, cpu_env, - cpu_V1, cpu_V0); - } else { - gen_helper_neon_qrshl_s64(cpu_V0, cpu_env, - cpu_V1, cpu_V0); - } - break; - default: - abort(); - } - neon_store_reg64(cpu_V0, rd + pass); - } - return 0; + /* 64-bit element instructions: handled by decodetree */ + return 1; } pairwise = 0; switch (op) { -- 2.20.1 Convert the Neon VHADD insns in the 3-reg-same group to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-5-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 2 ++ target/arm/translate-neon.inc.c | 24 ++++++++++++++++++++++++ target/arm/translate.c | 4 +--- 3 files changed, 27 insertions(+), 3 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 +VHADD_S_3s 1111 001 0 0 . .. .... .... 0000 . . . 0 .... @3same +VHADD_U_3s 1111 001 1 0 . .. .... .... 0000 . . . 0 .... @3same VQADD_S_3s 1111 001 0 0 . .. .... .... 0000 . . . 1 .... @3same VQADD_U_3s 1111 001 1 0 . .. .... .... 0000 . . . 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_64_ENV(VQSHL_S64, gen_helper_neon_qshl_s64) DO_3SAME_64_ENV(VQSHL_U64, gen_helper_neon_qshl_u64) DO_3SAME_64_ENV(VQRSHL_S64, gen_helper_neon_qrshl_s64) DO_3SAME_64_ENV(VQRSHL_U64, gen_helper_neon_qrshl_u64) + +#define DO_3SAME_32(INSN, FUNC) \ + 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) \ + { \ + static const GVecGen3 ops[4] = { \ + { .fni4 = gen_helper_neon_##FUNC##8 }, \ + { .fni4 = gen_helper_neon_##FUNC##16 }, \ + { .fni4 = gen_helper_neon_##FUNC##32 }, \ + { 0 }, \ + }; \ + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, oprsz, maxsz, &ops[vece]); \ + } \ + static bool trans_##INSN##_3s(DisasContext *s, arg_3same *a) \ + { \ + if (a->size > 2) { \ + return false; \ + } \ + return do_3same(s, a, gen_##INSN##_3s); \ + } + +DO_3SAME_32(VHADD_S, hadd_s) +DO_3SAME_32(VHADD_U, hadd_u) 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) case NEON_3R_VML: case NEON_3R_VSHL: case NEON_3R_SHA: + case NEON_3R_VHADD: /* Already handled by decodetree */ return 1; } @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) tmp2 = neon_load_reg(rm, pass); } switch (op) { - case NEON_3R_VHADD: - GEN_NEON_INTEGER_OP(hadd); - break; case NEON_3R_VRHADD: GEN_NEON_INTEGER_OP(rhadd); break; -- 2.20.1 Convert the Neon VABA and VABD insns in the 3-reg-same group to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-6-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 6 ++++++ target/arm/translate-neon.inc.c | 4 ++++ target/arm/translate.c | 22 ++-------------------- 3 files changed, 12 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 @@ 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 +VABD_S_3s 1111 001 0 0 . .. .... .... 0111 . . . 0 .... @3same +VABD_U_3s 1111 001 1 0 . .. .... .... 0111 . . . 0 .... @3same + +VABA_S_3s 1111 001 0 0 . .. .... .... 0111 . . . 1 .... @3same +VABA_U_3s 1111 001 1 0 . .. .... .... 0111 . . . 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_NO_SZ_3(VMUL, tcg_gen_gvec_mul) DO_3SAME_NO_SZ_3(VMLA, gen_gvec_mla) DO_3SAME_NO_SZ_3(VMLS, gen_gvec_mls) DO_3SAME_NO_SZ_3(VTST, gen_gvec_cmtst) +DO_3SAME_NO_SZ_3(VABD_S, gen_gvec_sabd) +DO_3SAME_NO_SZ_3(VABA_S, gen_gvec_saba) +DO_3SAME_NO_SZ_3(VABD_U, gen_gvec_uabd) +DO_3SAME_NO_SZ_3(VABA_U, gen_gvec_uaba) #define DO_3SAME_CMP(INSN, COND) \ static void gen_##INSN##_3s(unsigned vece, uint32_t rd_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) /* VQRDMLSH : handled by decodetree */ return 1; - case NEON_3R_VABD: - if (u) { - gen_gvec_uabd(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } else { - gen_gvec_sabd(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } - return 0; - - case NEON_3R_VABA: - if (u) { - gen_gvec_uaba(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } else { - gen_gvec_saba(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_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_VSHL: case NEON_3R_SHA: case NEON_3R_VHADD: + case NEON_3R_VABD: + case NEON_3R_VABA: /* Already handled by decodetree */ return 1; } -- 2.20.1 Convert the Neon VRHADD and VHSUB 3-reg-same insns to decodetree. (These are all the other insns in 3-reg-same which were using GEN_NEON_INTEGER_OP() and which are not pairwise or reversed-operands.) Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-7-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 6 ++++++ target/arm/translate-neon.inc.c | 4 ++++ target/arm/translate.c | 8 ++------ 3 files changed, 12 insertions(+), 6 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 @@ VHADD_U_3s 1111 001 1 0 . .. .... .... 0000 . . . 0 .... @3same VQADD_S_3s 1111 001 0 0 . .. .... .... 0000 . . . 1 .... @3same VQADD_U_3s 1111 001 1 0 . .. .... .... 0000 . . . 1 .... @3same +VRHADD_S_3s 1111 001 0 0 . .. .... .... 0001 . . . 0 .... @3same +VRHADD_U_3s 1111 001 1 0 . .. .... .... 0001 . . . 0 .... @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 +VHSUB_S_3s 1111 001 0 0 . .. .... .... 0010 . . . 0 .... @3same +VHSUB_U_3s 1111 001 1 0 . .. .... .... 0010 . . . 0 .... @3same + VQSUB_S_3s 1111 001 0 0 . .. .... .... 0010 . . . 1 .... @3same VQSUB_U_3s 1111 001 1 0 . .. .... .... 0010 . . . 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_64_ENV(VQRSHL_U64, gen_helper_neon_qrshl_u64) DO_3SAME_32(VHADD_S, hadd_s) DO_3SAME_32(VHADD_U, hadd_u) +DO_3SAME_32(VHSUB_S, hsub_s) +DO_3SAME_32(VHSUB_U, hsub_u) +DO_3SAME_32(VRHADD_S, rhadd_s) +DO_3SAME_32(VRHADD_U, rhadd_u) 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) case NEON_3R_VSHL: case NEON_3R_SHA: case NEON_3R_VHADD: + case NEON_3R_VRHADD: + case NEON_3R_VHSUB: case NEON_3R_VABD: case NEON_3R_VABA: /* Already handled by decodetree */ @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) tmp2 = neon_load_reg(rm, pass); } switch (op) { - case NEON_3R_VRHADD: - GEN_NEON_INTEGER_OP(rhadd); - break; - case NEON_3R_VHSUB: - GEN_NEON_INTEGER_OP(hsub); - break; case NEON_3R_VQSHL: GEN_NEON_INTEGER_OP_ENV(qshl); break; -- 2.20.1 Convert the VQSHL, VRSHL and VQRSHL insns in the 3-reg-same group to decodetree. We have already implemented the size==0b11 case of these insns; this commit handles the remaining sizes. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-8-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 30 ++++++++++++++++++----- target/arm/translate-neon.inc.c | 43 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 22 +++-------------- 3 files changed, 70 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 @@ VSHL_U_3s 1111 001 1 0 . .. .... .... 0100 . . . 0 .... @3same_rev @3same_64_rev .... ... . . . 11 .... .... .... . q:1 . . .... \ &3same vm=%vn_dp vn=%vm_dp vd=%vd_dp size=3 -VQSHL_S64_3s 1111 001 0 0 . .. .... .... 0100 . . . 1 .... @3same_64_rev -VQSHL_U64_3s 1111 001 1 0 . .. .... .... 0100 . . . 1 .... @3same_64_rev -VRSHL_S64_3s 1111 001 0 0 . .. .... .... 0101 . . . 0 .... @3same_64_rev -VRSHL_U64_3s 1111 001 1 0 . .. .... .... 0101 . . . 0 .... @3same_64_rev -VQRSHL_S64_3s 1111 001 0 0 . .. .... .... 0101 . . . 1 .... @3same_64_rev -VQRSHL_U64_3s 1111 001 1 0 . .. .... .... 0101 . . . 1 .... @3same_64_rev +{ + VQSHL_S64_3s 1111 001 0 0 . .. .... .... 0100 . . . 1 .... @3same_64_rev + VQSHL_S_3s 1111 001 0 0 . .. .... .... 0100 . . . 1 .... @3same_rev +} +{ + VQSHL_U64_3s 1111 001 1 0 . .. .... .... 0100 . . . 1 .... @3same_64_rev + VQSHL_U_3s 1111 001 1 0 . .. .... .... 0100 . . . 1 .... @3same_rev +} +{ + VRSHL_S64_3s 1111 001 0 0 . .. .... .... 0101 . . . 0 .... @3same_64_rev + VRSHL_S_3s 1111 001 0 0 . .. .... .... 0101 . . . 0 .... @3same_rev +} +{ + VRSHL_U64_3s 1111 001 1 0 . .. .... .... 0101 . . . 0 .... @3same_64_rev + VRSHL_U_3s 1111 001 1 0 . .. .... .... 0101 . . . 0 .... @3same_rev +} +{ + VQRSHL_S64_3s 1111 001 0 0 . .. .... .... 0101 . . . 1 .... @3same_64_rev + VQRSHL_S_3s 1111 001 0 0 . .. .... .... 0101 . . . 1 .... @3same_rev +} +{ + VQRSHL_U64_3s 1111 001 1 0 . .. .... .... 0101 . . . 1 .... @3same_64_rev + VQRSHL_U_3s 1111 001 1 0 . .. .... .... 0101 . . . 1 .... @3same_rev +} VMAX_S_3s 1111 001 0 0 . .. .... .... 0110 . . . 0 .... @3same VMAX_U_3s 1111 001 1 0 . .. .... .... 0110 . . . 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_64_ENV(VQRSHL_U64, gen_helper_neon_qrshl_u64) return do_3same(s, a, gen_##INSN##_3s); \ } +/* + * Some helper functions need to be passed the cpu_env. In order + * to use those with the gvec APIs like tcg_gen_gvec_3() we need + * to create wrapper functions whose prototype is a NeonGenTwoOpFn() + * and which call a NeonGenTwoOpEnvFn(). + */ +#define WRAP_ENV_FN(WRAPNAME, FUNC) \ + static void WRAPNAME(TCGv_i32 d, TCGv_i32 n, TCGv_i32 m) \ + { \ + FUNC(d, cpu_env, n, m); \ + } + +#define DO_3SAME_32_ENV(INSN, FUNC) \ + WRAP_ENV_FN(gen_##INSN##_tramp8, gen_helper_neon_##FUNC##8); \ + WRAP_ENV_FN(gen_##INSN##_tramp16, gen_helper_neon_##FUNC##16); \ + WRAP_ENV_FN(gen_##INSN##_tramp32, gen_helper_neon_##FUNC##32); \ + 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) \ + { \ + static const GVecGen3 ops[4] = { \ + { .fni4 = gen_##INSN##_tramp8 }, \ + { .fni4 = gen_##INSN##_tramp16 }, \ + { .fni4 = gen_##INSN##_tramp32 }, \ + { 0 }, \ + }; \ + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, oprsz, maxsz, &ops[vece]); \ + } \ + static bool trans_##INSN##_3s(DisasContext *s, arg_3same *a) \ + { \ + if (a->size > 2) { \ + return false; \ + } \ + return do_3same(s, a, gen_##INSN##_3s); \ + } + DO_3SAME_32(VHADD_S, hadd_s) DO_3SAME_32(VHADD_U, hadd_u) DO_3SAME_32(VHSUB_S, hsub_s) DO_3SAME_32(VHSUB_U, hsub_u) DO_3SAME_32(VRHADD_S, rhadd_s) DO_3SAME_32(VRHADD_U, rhadd_u) +DO_3SAME_32(VRSHL_S, rshl_s) +DO_3SAME_32(VRSHL_U, rshl_u) + +DO_3SAME_32_ENV(VQSHL_S, qshl_s) +DO_3SAME_32_ENV(VQSHL_U, qshl_u) +DO_3SAME_32_ENV(VQRSHL_S, qrshl_s) +DO_3SAME_32_ENV(VQRSHL_U, qrshl_u) 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) case NEON_3R_VHSUB: case NEON_3R_VABD: case NEON_3R_VABA: + case NEON_3R_VQSHL: + case NEON_3R_VRSHL: + case NEON_3R_VQRSHL: /* Already handled by decodetree */ return 1; } @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) } pairwise = 0; switch (op) { - case NEON_3R_VQSHL: - case NEON_3R_VRSHL: - case NEON_3R_VQRSHL: - { - int rtmp; - /* Shift instruction operands are reversed. */ - rtmp = rn; - rn = rm; - rm = rtmp; - } - break; case NEON_3R_VPADD_VQRDMLAH: case NEON_3R_VPMAX: case NEON_3R_VPMIN: @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) tmp2 = neon_load_reg(rm, pass); } switch (op) { - case NEON_3R_VQSHL: - GEN_NEON_INTEGER_OP_ENV(qshl); - break; - case NEON_3R_VRSHL: - GEN_NEON_INTEGER_OP(rshl); - break; - case NEON_3R_VQRSHL: - GEN_NEON_INTEGER_OP_ENV(qrshl); break; case NEON_3R_VPMAX: GEN_NEON_INTEGER_OP(pmax); -- 2.20.1 Convert the Neon integer VPMAX and VPMIN 3-reg-same insns to decodetree. These are 'pairwise' operations. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-9-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 9 +++++ target/arm/translate-neon.inc.c | 71 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 17 +------- 3 files changed, 82 insertions(+), 15 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_q0 .... ... . . . size:2 .... .... .... . 0 . . .... \ + &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp q=0 + VHADD_S_3s 1111 001 0 0 . .. .... .... 0000 . . . 0 .... @3same VHADD_U_3s 1111 001 1 0 . .. .... .... 0000 . . . 0 .... @3same VQADD_S_3s 1111 001 0 0 . .. .... .... 0000 . . . 1 .... @3same @@ -XXX,XX +XXX,XX @@ 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 +VPMAX_S_3s 1111 001 0 0 . .. .... .... 1010 . . . 0 .... @3same_q0 +VPMAX_U_3s 1111 001 1 0 . .. .... .... 1010 . . . 0 .... @3same_q0 + +VPMIN_S_3s 1111 001 0 0 . .. .... .... 1010 . . . 1 .... @3same_q0 +VPMIN_U_3s 1111 001 1 0 . .. .... .... 1010 . . . 1 .... @3same_q0 + VQRDMLAH_3s 1111 001 1 0 . .. .... .... 1011 ... 1 .... @3same SHA1_3s 1111 001 0 0 . optype:2 .... .... 1100 . 1 . 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 @@ DO_3SAME_32_ENV(VQSHL_S, qshl_s) DO_3SAME_32_ENV(VQSHL_U, qshl_u) DO_3SAME_32_ENV(VQRSHL_S, qrshl_s) DO_3SAME_32_ENV(VQRSHL_U, qrshl_u) + +static bool do_3same_pair(DisasContext *s, arg_3same *a, NeonGenTwoOpFn *fn) +{ + /* Operations handled pairwise 32 bits at a time */ + TCGv_i32 tmp, tmp2, tmp3; + + 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->size == 3) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + assert(a->q == 0); /* enforced by decode patterns */ + + /* + * Note that we have to be careful not to clobber the source operands + * in the "vm == vd" case by storing the result of the first pass too + * early. Since Q is 0 there are always just two passes, so instead + * of a complicated loop over each pass we just unroll. + */ + tmp = neon_load_reg(a->vn, 0); + tmp2 = neon_load_reg(a->vn, 1); + fn(tmp, tmp, tmp2); + tcg_temp_free_i32(tmp2); + + tmp3 = neon_load_reg(a->vm, 0); + tmp2 = neon_load_reg(a->vm, 1); + fn(tmp3, tmp3, tmp2); + tcg_temp_free_i32(tmp2); + + neon_store_reg(a->vd, 0, tmp); + neon_store_reg(a->vd, 1, tmp3); + return true; +} + +#define DO_3SAME_PAIR(INSN, func) \ + static bool trans_##INSN##_3s(DisasContext *s, arg_3same *a) \ + { \ + static NeonGenTwoOpFn * const fns[] = { \ + gen_helper_neon_##func##8, \ + gen_helper_neon_##func##16, \ + gen_helper_neon_##func##32, \ + }; \ + if (a->size > 2) { \ + return false; \ + } \ + return do_3same_pair(s, a, fns[a->size]); \ + } + +/* 32-bit pairwise ops end up the same as the elementwise versions. */ +#define gen_helper_neon_pmax_s32 tcg_gen_smax_i32 +#define gen_helper_neon_pmax_u32 tcg_gen_umax_i32 +#define gen_helper_neon_pmin_s32 tcg_gen_smin_i32 +#define gen_helper_neon_pmin_u32 tcg_gen_umin_i32 + +DO_3SAME_PAIR(VPMAX_S, pmax_s) +DO_3SAME_PAIR(VPMIN_S, pmin_s) +DO_3SAME_PAIR(VPMAX_U, pmax_u) +DO_3SAME_PAIR(VPMIN_U, pmin_u) 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 inline void gen_neon_rsb(int size, TCGv_i32 t0, TCGv_i32 t1) } } -/* 32-bit pairwise ops end up the same as the elementwise versions. */ -#define gen_helper_neon_pmax_s32 tcg_gen_smax_i32 -#define gen_helper_neon_pmax_u32 tcg_gen_umax_i32 -#define gen_helper_neon_pmin_s32 tcg_gen_smin_i32 -#define gen_helper_neon_pmin_u32 tcg_gen_umin_i32 - #define GEN_NEON_INTEGER_OP_ENV(name) do { \ switch ((size << 1) | u) { \ case 0: \ @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) case NEON_3R_VQSHL: case NEON_3R_VRSHL: case NEON_3R_VQRSHL: + case NEON_3R_VPMAX: + case NEON_3R_VPMIN: /* Already handled by decodetree */ return 1; } @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) pairwise = 0; switch (op) { case NEON_3R_VPADD_VQRDMLAH: - case NEON_3R_VPMAX: - case NEON_3R_VPMIN: pairwise = 1; break; case NEON_3R_FLOAT_ARITH: @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) tmp2 = neon_load_reg(rm, pass); } switch (op) { - break; - case NEON_3R_VPMAX: - GEN_NEON_INTEGER_OP(pmax); - break; - case NEON_3R_VPMIN: - GEN_NEON_INTEGER_OP(pmin); - break; case NEON_3R_VQDMULH_VQRDMULH: /* Multiply high. */ if (!u) { /* VQDMULH */ switch (size) { -- 2.20.1 Convert the Neon integer VPADD 3-reg-same insns to decodetree. These are 'pairwise' operations. (Note that VQRDMLAH, which shares the same primary opcode but has U=1, has already been converted.) Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-10-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 2 ++ target/arm/translate-neon.inc.c | 2 ++ target/arm/translate.c | 19 +------------------ 3 files changed, 5 insertions(+), 18 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 @@ VPMAX_U_3s 1111 001 1 0 . .. .... .... 1010 . . . 0 .... @3same_q0 VPMIN_S_3s 1111 001 0 0 . .. .... .... 1010 . . . 1 .... @3same_q0 VPMIN_U_3s 1111 001 1 0 . .. .... .... 1010 . . . 1 .... @3same_q0 +VPADD_3s 1111 001 0 0 . .. .... .... 1011 . . . 1 .... @3same_q0 + VQRDMLAH_3s 1111 001 1 0 . .. .... .... 1011 ... 1 .... @3same SHA1_3s 1111 001 0 0 . optype:2 .... .... 1100 . 1 . 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 do_3same_pair(DisasContext *s, arg_3same *a, NeonGenTwoOpFn *fn) #define gen_helper_neon_pmax_u32 tcg_gen_umax_i32 #define gen_helper_neon_pmin_s32 tcg_gen_smin_i32 #define gen_helper_neon_pmin_u32 tcg_gen_umin_i32 +#define gen_helper_neon_padd_u32 tcg_gen_add_i32 DO_3SAME_PAIR(VPMAX_S, pmax_s) DO_3SAME_PAIR(VPMIN_S, pmin_s) DO_3SAME_PAIR(VPMAX_U, pmax_u) DO_3SAME_PAIR(VPMIN_U, pmin_u) +DO_3SAME_PAIR(VPADD, padd_u) 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; } switch (op) { - case NEON_3R_VPADD_VQRDMLAH: - if (!u) { - break; /* VPADD */ - } - /* VQRDMLAH : handled by decodetree */ - return 1; - case NEON_3R_VFM_VQRDMLSH: if (!u) { /* VFM, VFMS */ @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) case NEON_3R_VQRSHL: case NEON_3R_VPMAX: case NEON_3R_VPMIN: + case NEON_3R_VPADD_VQRDMLAH: /* Already handled by decodetree */ return 1; } @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) } pairwise = 0; switch (op) { - case NEON_3R_VPADD_VQRDMLAH: - pairwise = 1; - break; case NEON_3R_FLOAT_ARITH: pairwise = (u && size < 2); /* if VPADD (float) */ break; @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) } } break; - case NEON_3R_VPADD_VQRDMLAH: - switch (size) { - case 0: gen_helper_neon_padd_u8(tmp, tmp, tmp2); break; - case 1: gen_helper_neon_padd_u16(tmp, tmp, tmp2); break; - case 2: tcg_gen_add_i32(tmp, tmp, tmp2); break; - default: abort(); - } - break; case NEON_3R_FLOAT_ARITH: /* Floating point arithmetic. */ { TCGv_ptr fpstatus = get_fpstatus_ptr(1); -- 2.20.1 Convert the Neon VQDMULH and VQRDMULH 3-reg-same insns to decodetree. These are the last integer operations in the 3-reg-same group. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-11-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 3 +++ target/arm/translate-neon.inc.c | 24 ++++++++++++++++++++++++ target/arm/translate.c | 24 +----------------------- 3 files changed, 28 insertions(+), 23 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 @@ VPMAX_U_3s 1111 001 1 0 . .. .... .... 1010 . . . 0 .... @3same_q0 VPMIN_S_3s 1111 001 0 0 . .. .... .... 1010 . . . 1 .... @3same_q0 VPMIN_U_3s 1111 001 1 0 . .. .... .... 1010 . . . 1 .... @3same_q0 +VQDMULH_3s 1111 001 0 0 . .. .... .... 1011 . . . 0 .... @3same +VQRDMULH_3s 1111 001 1 0 . .. .... .... 1011 . . . 0 .... @3same + VPADD_3s 1111 001 0 0 . .. .... .... 1011 . . . 1 .... @3same_q0 VQRDMLAH_3s 1111 001 1 0 . .. .... .... 1011 ... 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_PAIR(VPMIN_S, pmin_s) DO_3SAME_PAIR(VPMAX_U, pmax_u) DO_3SAME_PAIR(VPMIN_U, pmin_u) DO_3SAME_PAIR(VPADD, padd_u) + +#define DO_3SAME_VQDMULH(INSN, FUNC) \ + WRAP_ENV_FN(gen_##INSN##_tramp16, gen_helper_neon_##FUNC##_s16); \ + WRAP_ENV_FN(gen_##INSN##_tramp32, gen_helper_neon_##FUNC##_s32); \ + 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) \ + { \ + static const GVecGen3 ops[2] = { \ + { .fni4 = gen_##INSN##_tramp16 }, \ + { .fni4 = gen_##INSN##_tramp32 }, \ + }; \ + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, oprsz, maxsz, &ops[vece - 1]); \ + } \ + static bool trans_##INSN##_3s(DisasContext *s, arg_3same *a) \ + { \ + if (a->size != 1 && a->size != 2) { \ + return false; \ + } \ + return do_3same(s, a, gen_##INSN##_3s); \ + } + +DO_3SAME_VQDMULH(VQDMULH, qdmulh) +DO_3SAME_VQDMULH(VQRDMULH, qrdmulh) 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) case NEON_3R_VPMAX: case NEON_3R_VPMIN: case NEON_3R_VPADD_VQRDMLAH: + case NEON_3R_VQDMULH_VQRDMULH: /* Already handled by decodetree */ return 1; } @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) tmp2 = neon_load_reg(rm, pass); } switch (op) { - case NEON_3R_VQDMULH_VQRDMULH: /* Multiply high. */ - if (!u) { /* VQDMULH */ - switch (size) { - case 1: - gen_helper_neon_qdmulh_s16(tmp, cpu_env, tmp, tmp2); - break; - case 2: - gen_helper_neon_qdmulh_s32(tmp, cpu_env, tmp, tmp2); - break; - default: abort(); - } - } else { /* VQRDMULH */ - switch (size) { - case 1: - gen_helper_neon_qrdmulh_s16(tmp, cpu_env, tmp, tmp2); - break; - case 2: - gen_helper_neon_qrdmulh_s32(tmp, cpu_env, tmp, tmp2); - break; - default: abort(); - } - } - break; case NEON_3R_FLOAT_ARITH: /* Floating point arithmetic. */ { TCGv_ptr fpstatus = get_fpstatus_ptr(1); -- 2.20.1 Convert the Neon VADD, VSUB, VABD 3-reg-same insns to decodetree. We already have gvec helpers for addition and subtraction, but must add one for fabd. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-12-peter.maydell@linaro.org --- target/arm/helper.h | 3 ++- target/arm/neon-dp.decode | 8 ++++++++ target/arm/neon_helper.c | 7 ------- target/arm/translate-neon.inc.c | 28 ++++++++++++++++++++++++++++ target/arm/translate.c | 10 +++------- target/arm/vec_helper.c | 7 +++++++ 6 files changed, 48 insertions(+), 15 deletions(-) diff --git a/target/arm/helper.h b/target/arm/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_2(neon_qneg_s16, TCG_CALL_NO_RWG, i32, env, i32) DEF_HELPER_FLAGS_2(neon_qneg_s32, TCG_CALL_NO_RWG, i32, env, i32) DEF_HELPER_FLAGS_2(neon_qneg_s64, TCG_CALL_NO_RWG, i64, env, i64) -DEF_HELPER_3(neon_abd_f32, i32, i32, i32, ptr) DEF_HELPER_3(neon_ceq_f32, i32, i32, i32, ptr) DEF_HELPER_3(neon_cge_f32, i32, i32, i32, ptr) DEF_HELPER_3(neon_cgt_f32, i32, i32, i32, ptr) @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_5(gvec_fmul_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(gvec_fmul_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(gvec_fmul_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_fabd_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) + DEF_HELPER_FLAGS_5(gvec_ftsmul_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(gvec_ftsmul_s, TCG_CALL_NO_RWG, 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_q0 .... ... . . . size:2 .... .... .... . 0 . . .... \ &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp q=0 +# For FP insns the high bit of 'size' is used as part of opcode decode +@3same_fp .... ... . . . . size:1 .... .... .... . q:1 . . .... \ + &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp + VHADD_S_3s 1111 001 0 0 . .. .... .... 0000 . . . 0 .... @3same VHADD_U_3s 1111 001 1 0 . .. .... .... 0000 . . . 0 .... @3same VQADD_S_3s 1111 001 0 0 . .. .... .... 0000 . . . 1 .... @3same @@ -XXX,XX +XXX,XX @@ SHA256SU1_3s 1111 001 1 0 . 10 .... .... 1100 . 1 . 0 .... \ vm=%vm_dp vn=%vn_dp vd=%vd_dp VQRDMLSH_3s 1111 001 1 0 . .. .... .... 1100 ... 1 .... @3same + +VADD_fp_3s 1111 001 0 0 . 0 . .... .... 1101 ... 0 .... @3same_fp +VSUB_fp_3s 1111 001 0 0 . 1 . .... .... 1101 ... 0 .... @3same_fp +VABD_fp_3s 1111 001 1 0 . 1 . .... .... 1101 ... 0 .... @3same_fp diff --git a/target/arm/neon_helper.c b/target/arm/neon_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon_helper.c +++ b/target/arm/neon_helper.c @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(neon_qneg_s64)(CPUARMState *env, uint64_t x) } /* NEON Float helpers. */ -uint32_t HELPER(neon_abd_f32)(uint32_t a, uint32_t b, void *fpstp) -{ - float_status *fpst = fpstp; - float32 f0 = make_float32(a); - float32 f1 = make_float32(b); - return float32_val(float32_abs(float32_sub(f0, f1, fpst))); -} /* Floating point comparisons produce an integer result. * Note that EQ doesn't signal InvalidOp for QNaNs but GE and GT do. 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_PAIR(VPADD, padd_u) DO_3SAME_VQDMULH(VQDMULH, qdmulh) DO_3SAME_VQDMULH(VQRDMULH, qrdmulh) + +/* + * For all the functions using this macro, size == 1 means fp16, + * which is an architecture extension we don't implement yet. + */ +#define DO_3S_FP_GVEC(INSN,FUNC) \ + 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) \ + { \ + TCGv_ptr fpst = get_fpstatus_ptr(1); \ + tcg_gen_gvec_3_ptr(rd_ofs, rn_ofs, rm_ofs, fpst, \ + oprsz, maxsz, 0, FUNC); \ + tcg_temp_free_ptr(fpst); \ + } \ + static bool trans_##INSN##_fp_3s(DisasContext *s, arg_3same *a) \ + { \ + if (a->size != 0) { \ + /* TODO fp16 support */ \ + return false; \ + } \ + return do_3same(s, a, gen_##INSN##_3s); \ + } + + +DO_3S_FP_GVEC(VADD, gen_helper_gvec_fadd_s) +DO_3S_FP_GVEC(VSUB, gen_helper_gvec_fsub_s) +DO_3S_FP_GVEC(VABD, gen_helper_gvec_fabd_s) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) switch (op) { case NEON_3R_FLOAT_ARITH: pairwise = (u && size < 2); /* if VPADD (float) */ + if (!pairwise) { + return 1; /* handled by decodetree */ + } break; case NEON_3R_FLOAT_MINMAX: pairwise = u; /* if VPMIN/VPMAX (float) */ @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) { TCGv_ptr fpstatus = get_fpstatus_ptr(1); switch ((u << 2) | size) { - case 0: /* VADD */ case 4: /* VPADD */ gen_helper_vfp_adds(tmp, tmp, tmp2, fpstatus); break; - case 2: /* VSUB */ - gen_helper_vfp_subs(tmp, tmp, tmp2, fpstatus); - break; - case 6: /* VABD */ - gen_helper_neon_abd_f32(tmp, tmp, tmp2, fpstatus); - break; default: abort(); } diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vec_helper.c +++ b/target/arm/vec_helper.c @@ -XXX,XX +XXX,XX @@ static float64 float64_ftsmul(float64 op1, uint64_t op2, float_status *stat) return result; } +static float32 float32_abd(float32 op1, float32 op2, float_status *stat) +{ + return float32_abs(float32_sub(op1, op2, stat)); +} + #define DO_3OP(NAME, FUNC, TYPE) \ void HELPER(NAME)(void *vd, void *vn, void *vm, void *stat, uint32_t desc) \ { \ @@ -XXX,XX +XXX,XX @@ DO_3OP(gvec_ftsmul_h, float16_ftsmul, float16) DO_3OP(gvec_ftsmul_s, float32_ftsmul, float32) DO_3OP(gvec_ftsmul_d, float64_ftsmul, float64) +DO_3OP(gvec_fabd_s, float32_abd, float32) + #ifdef TARGET_AARCH64 DO_3OP(gvec_recps_h, helper_recpsf_f16, float16) -- 2.20.1 Convert the Neon float VPMIN, VPMAX and VPADD 3-reg-same insns to decodetree. These are the only remaining 'pairwise' operations, so we can delete the pairwise-specific bits of the old decoder's for-each-element loop now. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-13-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 5 +++ target/arm/translate-neon.inc.c | 63 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 63 +++++---------------------------- 3 files changed, 76 insertions(+), 55 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 @@ # For FP insns the high bit of 'size' is used as part of opcode decode @3same_fp .... ... . . . . size:1 .... .... .... . q:1 . . .... \ &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp +@3same_fp_q0 .... ... . . . . size:1 .... .... .... . 0 . . .... \ + &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp q=0 VHADD_S_3s 1111 001 0 0 . .. .... .... 0000 . . . 0 .... @3same VHADD_U_3s 1111 001 1 0 . .. .... .... 0000 . . . 0 .... @3same @@ -XXX,XX +XXX,XX @@ VQRDMLSH_3s 1111 001 1 0 . .. .... .... 1100 ... 1 .... @3same VADD_fp_3s 1111 001 0 0 . 0 . .... .... 1101 ... 0 .... @3same_fp VSUB_fp_3s 1111 001 0 0 . 1 . .... .... 1101 ... 0 .... @3same_fp +VPADD_fp_3s 1111 001 1 0 . 0 . .... .... 1101 ... 0 .... @3same_fp_q0 VABD_fp_3s 1111 001 1 0 . 1 . .... .... 1101 ... 0 .... @3same_fp +VPMAX_fp_3s 1111 001 1 0 . 0 . .... .... 1111 ... 0 .... @3same_fp_q0 +VPMIN_fp_3s 1111 001 1 0 . 1 . .... .... 1111 ... 0 .... @3same_fp_q0 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_VQDMULH(VQRDMULH, qrdmulh) DO_3S_FP_GVEC(VADD, gen_helper_gvec_fadd_s) DO_3S_FP_GVEC(VSUB, gen_helper_gvec_fsub_s) DO_3S_FP_GVEC(VABD, gen_helper_gvec_fabd_s) + +static bool do_3same_fp_pair(DisasContext *s, arg_3same *a, VFPGen3OpSPFn *fn) +{ + /* FP operations handled pairwise 32 bits at a time */ + TCGv_i32 tmp, tmp2, tmp3; + TCGv_ptr fpstatus; + + 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 (!vfp_access_check(s)) { + return true; + } + + assert(a->q == 0); /* enforced by decode patterns */ + + /* + * Note that we have to be careful not to clobber the source operands + * in the "vm == vd" case by storing the result of the first pass too + * early. Since Q is 0 there are always just two passes, so instead + * of a complicated loop over each pass we just unroll. + */ + fpstatus = get_fpstatus_ptr(1); + tmp = neon_load_reg(a->vn, 0); + tmp2 = neon_load_reg(a->vn, 1); + fn(tmp, tmp, tmp2, fpstatus); + tcg_temp_free_i32(tmp2); + + tmp3 = neon_load_reg(a->vm, 0); + tmp2 = neon_load_reg(a->vm, 1); + fn(tmp3, tmp3, tmp2, fpstatus); + tcg_temp_free_i32(tmp2); + tcg_temp_free_ptr(fpstatus); + + neon_store_reg(a->vd, 0, tmp); + neon_store_reg(a->vd, 1, tmp3); + return true; +} + +/* + * For all the functions using this macro, size == 1 means fp16, + * which is an architecture extension we don't implement yet. + */ +#define DO_3S_FP_PAIR(INSN,FUNC) \ + static bool trans_##INSN##_fp_3s(DisasContext *s, arg_3same *a) \ + { \ + if (a->size != 0) { \ + /* TODO fp16 support */ \ + return false; \ + } \ + return do_3same_fp_pair(s, a, FUNC); \ + } + +DO_3S_FP_PAIR(VPADD, gen_helper_vfp_adds) +DO_3S_FP_PAIR(VPMAX, gen_helper_vfp_maxs) +DO_3S_FP_PAIR(VPMIN, gen_helper_vfp_mins) 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) int shift; int pass; int count; - int pairwise; int u; int vec_size; uint32_t imm; @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) case NEON_3R_VPMIN: case NEON_3R_VPADD_VQRDMLAH: case NEON_3R_VQDMULH_VQRDMULH: + case NEON_3R_FLOAT_ARITH: /* Already handled by decodetree */ return 1; } @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) /* 64-bit element instructions: handled by decodetree */ return 1; } - pairwise = 0; switch (op) { - case NEON_3R_FLOAT_ARITH: - pairwise = (u && size < 2); /* if VPADD (float) */ - if (!pairwise) { - return 1; /* handled by decodetree */ - } - break; case NEON_3R_FLOAT_MINMAX: - pairwise = u; /* if VPMIN/VPMAX (float) */ + if (u) { + return 1; /* VPMIN/VPMAX handled by decodetree */ + } break; case NEON_3R_FLOAT_CMP: if (!u && size) { @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) break; } - if (pairwise && q) { - /* All the pairwise insns UNDEF if Q is set */ - return 1; - } - for (pass = 0; pass < (q ? 4 : 2); pass++) { - if (pairwise) { - /* Pairwise. */ - if (pass < 1) { - tmp = neon_load_reg(rn, 0); - tmp2 = neon_load_reg(rn, 1); - } else { - tmp = neon_load_reg(rm, 0); - tmp2 = neon_load_reg(rm, 1); - } - } else { - /* Elementwise. */ - tmp = neon_load_reg(rn, pass); - tmp2 = neon_load_reg(rm, pass); - } + /* Elementwise. */ + tmp = neon_load_reg(rn, pass); + tmp2 = neon_load_reg(rm, pass); switch (op) { - case NEON_3R_FLOAT_ARITH: /* Floating point arithmetic. */ - { - TCGv_ptr fpstatus = get_fpstatus_ptr(1); - switch ((u << 2) | size) { - case 4: /* VPADD */ - gen_helper_vfp_adds(tmp, tmp, tmp2, fpstatus); - break; - default: - abort(); - } - tcg_temp_free_ptr(fpstatus); - break; - } case NEON_3R_FLOAT_MULTIPLY: { TCGv_ptr fpstatus = get_fpstatus_ptr(1); @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) } tcg_temp_free_i32(tmp2); - /* Save the result. For elementwise operations we can put it - straight into the destination register. For pairwise operations - we have to be careful to avoid clobbering the source operands. */ - if (pairwise && rd == rm) { - neon_store_scratch(pass, tmp); - } else { - neon_store_reg(rd, pass, tmp); - } + neon_store_reg(rd, pass, tmp); } /* for pass */ - if (pairwise && rd == rm) { - for (pass = 0; pass < (q ? 4 : 2); pass++) { - tmp = neon_load_scratch(pass); - neon_store_reg(rd, pass, tmp); - } - } /* End of 3 register same size operations. */ } else if (insn & (1 << 4)) { if ((insn & 0x00380080) != 0) { -- 2.20.1 Convert the Neon integer VMUL, VMLA, and VMLS 3-reg-same inssn to decodetree. We don't have a gvec helper for multiply-accumulate, so VMLA and VMLS need a loop function do_3same_fp(). This takes a reads_vd parameter to do_3same_fp() which tells it to load the old value into vd before calling the callback function, in the same way that the do_vfp_3op_sp() and do_vfp_3op_dp() functions in translate-vfp.inc.c work. (The only uses in this patch pass reads_vd == true, but later commits will use reads_vd == false.) This conversion fixes in passing an underdecoding for VMUL (originally reported by Fredrik Strupe <fredrik@strupe.net>): bit 1 of the 'size' field must be 0. The old decoder didn't enforce this, but the decodetree pattern does. The gen_VMLA_fp_reg() function performs the addition operation with the operands in the opposite order to the old decoder: since Neon sets 'default NaN mode' float32_add operations are commutative so there is no behaviour difference, but putting them this way around matches the Arm ARM pseudocode and the required operation order for the subtraction in gen_VMLS_fp_reg(). Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-14-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 3 ++ target/arm/translate-neon.inc.c | 81 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 17 +------ 3 files changed, 85 insertions(+), 16 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 @@ VADD_fp_3s 1111 001 0 0 . 0 . .... .... 1101 ... 0 .... @3same_fp VSUB_fp_3s 1111 001 0 0 . 1 . .... .... 1101 ... 0 .... @3same_fp VPADD_fp_3s 1111 001 1 0 . 0 . .... .... 1101 ... 0 .... @3same_fp_q0 VABD_fp_3s 1111 001 1 0 . 1 . .... .... 1101 ... 0 .... @3same_fp +VMLA_fp_3s 1111 001 0 0 . 0 . .... .... 1101 ... 1 .... @3same_fp +VMLS_fp_3s 1111 001 0 0 . 1 . .... .... 1101 ... 1 .... @3same_fp +VMUL_fp_3s 1111 001 1 0 . 0 . .... .... 1101 ... 1 .... @3same_fp VPMAX_fp_3s 1111 001 1 0 . 0 . .... .... 1111 ... 0 .... @3same_fp_q0 VPMIN_fp_3s 1111 001 1 0 . 1 . .... .... 1111 ... 0 .... @3same_fp_q0 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_PAIR(VPADD, padd_u) DO_3SAME_VQDMULH(VQDMULH, qdmulh) DO_3SAME_VQDMULH(VQRDMULH, qrdmulh) +static bool do_3same_fp(DisasContext *s, arg_3same *a, VFPGen3OpSPFn *fn, + bool reads_vd) +{ + /* + * FP operations handled elementwise 32 bits at a time. + * If reads_vd is true then the old value of Vd will be + * loaded before calling the callback function. This is + * used for multiply-accumulate type operations. + */ + TCGv_i32 tmp, tmp2; + int pass; + + 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; + } + + TCGv_ptr fpstatus = get_fpstatus_ptr(1); + for (pass = 0; pass < (a->q ? 4 : 2); pass++) { + tmp = neon_load_reg(a->vn, pass); + tmp2 = neon_load_reg(a->vm, pass); + if (reads_vd) { + TCGv_i32 tmp_rd = neon_load_reg(a->vd, pass); + fn(tmp_rd, tmp, tmp2, fpstatus); + neon_store_reg(a->vd, pass, tmp_rd); + tcg_temp_free_i32(tmp); + } else { + fn(tmp, tmp, tmp2, fpstatus); + neon_store_reg(a->vd, pass, tmp); + } + tcg_temp_free_i32(tmp2); + } + tcg_temp_free_ptr(fpstatus); + return true; +} + /* * For all the functions using this macro, size == 1 means fp16, * which is an architecture extension we don't implement yet. @@ -XXX,XX +XXX,XX @@ DO_3SAME_VQDMULH(VQRDMULH, qrdmulh) DO_3S_FP_GVEC(VADD, gen_helper_gvec_fadd_s) DO_3S_FP_GVEC(VSUB, gen_helper_gvec_fsub_s) DO_3S_FP_GVEC(VABD, gen_helper_gvec_fabd_s) +DO_3S_FP_GVEC(VMUL, gen_helper_gvec_fmul_s) + +/* + * For all the functions using this macro, size == 1 means fp16, + * which is an architecture extension we don't implement yet. + */ +#define DO_3S_FP(INSN,FUNC,READS_VD) \ + static bool trans_##INSN##_fp_3s(DisasContext *s, arg_3same *a) \ + { \ + if (a->size != 0) { \ + /* TODO fp16 support */ \ + return false; \ + } \ + return do_3same_fp(s, a, FUNC, READS_VD); \ + } + +static void gen_VMLA_fp_3s(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, + TCGv_ptr fpstatus) +{ + gen_helper_vfp_muls(vn, vn, vm, fpstatus); + gen_helper_vfp_adds(vd, vd, vn, fpstatus); +} + +static void gen_VMLS_fp_3s(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, + TCGv_ptr fpstatus) +{ + gen_helper_vfp_muls(vn, vn, vm, fpstatus); + gen_helper_vfp_subs(vd, vd, vn, fpstatus); +} + +DO_3S_FP(VMLA, gen_VMLA_fp_3s, true) +DO_3S_FP(VMLS, gen_VMLS_fp_3s, true) static bool do_3same_fp_pair(DisasContext *s, arg_3same *a, VFPGen3OpSPFn *fn) { 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) case NEON_3R_VPADD_VQRDMLAH: case NEON_3R_VQDMULH_VQRDMULH: case NEON_3R_FLOAT_ARITH: + case NEON_3R_FLOAT_MULTIPLY: /* Already handled by decodetree */ return 1; } @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) tmp = neon_load_reg(rn, pass); tmp2 = neon_load_reg(rm, pass); switch (op) { - case NEON_3R_FLOAT_MULTIPLY: - { - TCGv_ptr fpstatus = get_fpstatus_ptr(1); - gen_helper_vfp_muls(tmp, tmp, tmp2, fpstatus); - if (!u) { - tcg_temp_free_i32(tmp2); - tmp2 = neon_load_reg(rd, pass); - if (size == 0) { - gen_helper_vfp_adds(tmp, tmp, tmp2, fpstatus); - } else { - gen_helper_vfp_subs(tmp, tmp2, tmp, fpstatus); - } - } - tcg_temp_free_ptr(fpstatus); - break; - } case NEON_3R_FLOAT_CMP: { TCGv_ptr fpstatus = get_fpstatus_ptr(1); -- 2.20.1 Convert the Neon integer 3-reg-same compare insns VCGE, VCGT, VCEQ, VACGE and VACGT to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-15-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 5 +++++ target/arm/translate-neon.inc.c | 6 +++++ target/arm/translate.c | 39 ++------------------------------- 3 files changed, 13 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 @@ VABD_fp_3s 1111 001 1 0 . 1 . .... .... 1101 ... 0 .... @3same_fp VMLA_fp_3s 1111 001 0 0 . 0 . .... .... 1101 ... 1 .... @3same_fp VMLS_fp_3s 1111 001 0 0 . 1 . .... .... 1101 ... 1 .... @3same_fp VMUL_fp_3s 1111 001 1 0 . 0 . .... .... 1101 ... 1 .... @3same_fp +VCEQ_fp_3s 1111 001 0 0 . 0 . .... .... 1110 ... 0 .... @3same_fp +VCGE_fp_3s 1111 001 1 0 . 0 . .... .... 1110 ... 0 .... @3same_fp +VACGE_fp_3s 1111 001 1 0 . 0 . .... .... 1110 ... 1 .... @3same_fp +VCGT_fp_3s 1111 001 1 0 . 1 . .... .... 1110 ... 0 .... @3same_fp +VACGT_fp_3s 1111 001 1 0 . 1 . .... .... 1110 ... 1 .... @3same_fp VPMAX_fp_3s 1111 001 1 0 . 0 . .... .... 1111 ... 0 .... @3same_fp_q0 VPMIN_fp_3s 1111 001 1 0 . 1 . .... .... 1111 ... 0 .... @3same_fp_q0 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_3S_FP_GVEC(VMUL, gen_helper_gvec_fmul_s) return do_3same_fp(s, a, FUNC, READS_VD); \ } +DO_3S_FP(VCEQ, gen_helper_neon_ceq_f32, false) +DO_3S_FP(VCGE, gen_helper_neon_cge_f32, false) +DO_3S_FP(VCGT, gen_helper_neon_cgt_f32, false) +DO_3S_FP(VACGE, gen_helper_neon_acge_f32, false) +DO_3S_FP(VACGT, gen_helper_neon_acgt_f32, false) + static void gen_VMLA_fp_3s(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, TCGv_ptr fpstatus) { 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) case NEON_3R_VQDMULH_VQRDMULH: case NEON_3R_FLOAT_ARITH: case NEON_3R_FLOAT_MULTIPLY: + case NEON_3R_FLOAT_CMP: + case NEON_3R_FLOAT_ACMP: /* Already handled by decodetree */ return 1; } @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) return 1; /* VPMIN/VPMAX handled by decodetree */ } break; - case NEON_3R_FLOAT_CMP: - if (!u && size) { - /* no encoding for U=0 C=1x */ - return 1; - } - break; - case NEON_3R_FLOAT_ACMP: - if (!u) { - return 1; - } - break; case NEON_3R_FLOAT_MISC: /* VMAXNM/VMINNM in ARMv8 */ if (u && !arm_dc_feature(s, ARM_FEATURE_V8)) { @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) tmp = neon_load_reg(rn, pass); tmp2 = neon_load_reg(rm, pass); switch (op) { - case NEON_3R_FLOAT_CMP: - { - TCGv_ptr fpstatus = get_fpstatus_ptr(1); - if (!u) { - gen_helper_neon_ceq_f32(tmp, tmp, tmp2, fpstatus); - } else { - if (size == 0) { - gen_helper_neon_cge_f32(tmp, tmp, tmp2, fpstatus); - } else { - gen_helper_neon_cgt_f32(tmp, tmp, tmp2, fpstatus); - } - } - tcg_temp_free_ptr(fpstatus); - break; - } - case NEON_3R_FLOAT_ACMP: - { - TCGv_ptr fpstatus = get_fpstatus_ptr(1); - if (size == 0) { - gen_helper_neon_acge_f32(tmp, tmp, tmp2, fpstatus); - } else { - gen_helper_neon_acgt_f32(tmp, tmp, tmp2, fpstatus); - } - tcg_temp_free_ptr(fpstatus); - break; - } case NEON_3R_FLOAT_MINMAX: { TCGv_ptr fpstatus = get_fpstatus_ptr(1); -- 2.20.1 The usual location for the env argument in the argument list of a TCG helper is immediately after the return-value argument. recps_f32 and rsqrts_f32 differ in that they put it at the end. Move the env argument to its usual place; this will allow us to more easily use these helper functions with the gvec APIs. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-16-peter.maydell@linaro.org --- target/arm/helper.h | 4 ++-- target/arm/translate.c | 4 ++-- target/arm/vfp_helper.c | 4 ++-- 3 files changed, 6 insertions(+), 6 deletions(-) diff --git a/target/arm/helper.h b/target/arm/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(vfp_fcvt_f64_to_f16, TCG_CALL_NO_RWG, f16, f64, ptr, i32) DEF_HELPER_4(vfp_muladdd, f64, f64, f64, f64, ptr) DEF_HELPER_4(vfp_muladds, f32, f32, f32, f32, ptr) -DEF_HELPER_3(recps_f32, f32, f32, f32, env) -DEF_HELPER_3(rsqrts_f32, f32, f32, f32, env) +DEF_HELPER_3(recps_f32, f32, env, f32, f32) +DEF_HELPER_3(rsqrts_f32, f32, env, f32, f32) DEF_HELPER_FLAGS_2(recpe_f16, TCG_CALL_NO_RWG, f16, f16, ptr) DEF_HELPER_FLAGS_2(recpe_f32, TCG_CALL_NO_RWG, f32, f32, ptr) DEF_HELPER_FLAGS_2(recpe_f64, TCG_CALL_NO_RWG, f64, f64, ptr) 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) tcg_temp_free_ptr(fpstatus); } else { if (size == 0) { - gen_helper_recps_f32(tmp, tmp, tmp2, cpu_env); + gen_helper_recps_f32(tmp, cpu_env, tmp, tmp2); } else { - gen_helper_rsqrts_f32(tmp, tmp, tmp2, cpu_env); + gen_helper_rsqrts_f32(tmp, cpu_env, tmp, tmp2); } } break; diff --git a/target/arm/vfp_helper.c b/target/arm/vfp_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vfp_helper.c +++ b/target/arm/vfp_helper.c @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(vfp_fcvt_f64_to_f16)(float64 a, void *fpstp, uint32_t ahp_mode) #define float32_three make_float32(0x40400000) #define float32_one_point_five make_float32(0x3fc00000) -float32 HELPER(recps_f32)(float32 a, float32 b, CPUARMState *env) +float32 HELPER(recps_f32)(CPUARMState *env, float32 a, float32 b) { float_status *s = &env->vfp.standard_fp_status; if ((float32_is_infinity(a) && float32_is_zero_or_denormal(b)) || @@ -XXX,XX +XXX,XX @@ float32 HELPER(recps_f32)(float32 a, float32 b, CPUARMState *env) return float32_sub(float32_two, float32_mul(a, b, s), s); } -float32 HELPER(rsqrts_f32)(float32 a, float32 b, CPUARMState *env) +float32 HELPER(rsqrts_f32)(CPUARMState *env, float32 a, float32 b) { float_status *s = &env->vfp.standard_fp_status; float32 product; -- 2.20.1 Convert the Neon fp VMAX/VMIN/VMAXNM/VMINNM/VRECPS/VRSQRTS 3-reg-same insns to decodetree. (These are all the remaining non-accumulation instructions in this group.) Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-17-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 6 +++ target/arm/translate-neon.inc.c | 70 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 42 +------------------- 3 files changed, 78 insertions(+), 40 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 @@ VCGE_fp_3s 1111 001 1 0 . 0 . .... .... 1110 ... 0 .... @3same_fp VACGE_fp_3s 1111 001 1 0 . 0 . .... .... 1110 ... 1 .... @3same_fp VCGT_fp_3s 1111 001 1 0 . 1 . .... .... 1110 ... 0 .... @3same_fp VACGT_fp_3s 1111 001 1 0 . 1 . .... .... 1110 ... 1 .... @3same_fp +VMAX_fp_3s 1111 001 0 0 . 0 . .... .... 1111 ... 0 .... @3same_fp +VMIN_fp_3s 1111 001 0 0 . 1 . .... .... 1111 ... 0 .... @3same_fp VPMAX_fp_3s 1111 001 1 0 . 0 . .... .... 1111 ... 0 .... @3same_fp_q0 VPMIN_fp_3s 1111 001 1 0 . 1 . .... .... 1111 ... 0 .... @3same_fp_q0 +VRECPS_fp_3s 1111 001 0 0 . 0 . .... .... 1111 ... 1 .... @3same_fp +VRSQRTS_fp_3s 1111 001 0 0 . 1 . .... .... 1111 ... 1 .... @3same_fp +VMAXNM_fp_3s 1111 001 1 0 . 0 . .... .... 1111 ... 1 .... @3same_fp +VMINNM_fp_3s 1111 001 1 0 . 1 . .... .... 1111 ... 1 .... @3same_fp 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_3S_FP(VCGE, gen_helper_neon_cge_f32, false) DO_3S_FP(VCGT, gen_helper_neon_cgt_f32, false) DO_3S_FP(VACGE, gen_helper_neon_acge_f32, false) DO_3S_FP(VACGT, gen_helper_neon_acgt_f32, false) +DO_3S_FP(VMAX, gen_helper_vfp_maxs, false) +DO_3S_FP(VMIN, gen_helper_vfp_mins, false) static void gen_VMLA_fp_3s(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, TCGv_ptr fpstatus) @@ -XXX,XX +XXX,XX @@ static void gen_VMLS_fp_3s(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, DO_3S_FP(VMLA, gen_VMLA_fp_3s, true) DO_3S_FP(VMLS, gen_VMLS_fp_3s, true) +static bool trans_VMAXNM_fp_3s(DisasContext *s, arg_3same *a) +{ + if (!arm_dc_feature(s, ARM_FEATURE_V8)) { + return false; + } + + if (a->size != 0) { + /* TODO fp16 support */ + return false; + } + + return do_3same_fp(s, a, gen_helper_vfp_maxnums, false); +} + +static bool trans_VMINNM_fp_3s(DisasContext *s, arg_3same *a) +{ + if (!arm_dc_feature(s, ARM_FEATURE_V8)) { + return false; + } + + if (a->size != 0) { + /* TODO fp16 support */ + return false; + } + + return do_3same_fp(s, a, gen_helper_vfp_minnums, false); +} + +WRAP_ENV_FN(gen_VRECPS_tramp, gen_helper_recps_f32) + +static void gen_VRECPS_fp_3s(unsigned vece, uint32_t rd_ofs, + uint32_t rn_ofs, uint32_t rm_ofs, + uint32_t oprsz, uint32_t maxsz) +{ + static const GVecGen3 ops = { .fni4 = gen_VRECPS_tramp }; + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, oprsz, maxsz, &ops); +} + +static bool trans_VRECPS_fp_3s(DisasContext *s, arg_3same *a) +{ + if (a->size != 0) { + /* TODO fp16 support */ + return false; + } + + return do_3same(s, a, gen_VRECPS_fp_3s); +} + +WRAP_ENV_FN(gen_VRSQRTS_tramp, gen_helper_rsqrts_f32) + +static void gen_VRSQRTS_fp_3s(unsigned vece, uint32_t rd_ofs, + uint32_t rn_ofs, uint32_t rm_ofs, + uint32_t oprsz, uint32_t maxsz) +{ + static const GVecGen3 ops = { .fni4 = gen_VRSQRTS_tramp }; + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, oprsz, maxsz, &ops); +} + +static bool trans_VRSQRTS_fp_3s(DisasContext *s, arg_3same *a) +{ + if (a->size != 0) { + /* TODO fp16 support */ + return false; + } + + return do_3same(s, a, gen_VRSQRTS_fp_3s); +} + static bool do_3same_fp_pair(DisasContext *s, arg_3same *a, VFPGen3OpSPFn *fn) { /* FP operations handled pairwise 32 bits at a time */ 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) case NEON_3R_FLOAT_MULTIPLY: case NEON_3R_FLOAT_CMP: case NEON_3R_FLOAT_ACMP: + case NEON_3R_FLOAT_MINMAX: + case NEON_3R_FLOAT_MISC: /* Already handled by decodetree */ return 1; } @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) return 1; } switch (op) { - case NEON_3R_FLOAT_MINMAX: - if (u) { - return 1; /* VPMIN/VPMAX handled by decodetree */ - } - break; - case NEON_3R_FLOAT_MISC: - /* VMAXNM/VMINNM in ARMv8 */ - if (u && !arm_dc_feature(s, ARM_FEATURE_V8)) { - return 1; - } - break; case NEON_3R_VFM_VQRDMLSH: if (!dc_isar_feature(aa32_simdfmac, s)) { return 1; @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) tmp = neon_load_reg(rn, pass); tmp2 = neon_load_reg(rm, pass); switch (op) { - case NEON_3R_FLOAT_MINMAX: - { - TCGv_ptr fpstatus = get_fpstatus_ptr(1); - if (size == 0) { - gen_helper_vfp_maxs(tmp, tmp, tmp2, fpstatus); - } else { - gen_helper_vfp_mins(tmp, tmp, tmp2, fpstatus); - } - tcg_temp_free_ptr(fpstatus); - break; - } - case NEON_3R_FLOAT_MISC: - if (u) { - /* VMAXNM/VMINNM */ - TCGv_ptr fpstatus = get_fpstatus_ptr(1); - if (size == 0) { - gen_helper_vfp_maxnums(tmp, tmp, tmp2, fpstatus); - } else { - gen_helper_vfp_minnums(tmp, tmp, tmp2, fpstatus); - } - tcg_temp_free_ptr(fpstatus); - } else { - if (size == 0) { - gen_helper_recps_f32(tmp, cpu_env, tmp, tmp2); - } else { - gen_helper_rsqrts_f32(tmp, cpu_env, tmp, tmp2); - } - } - break; case NEON_3R_VFM_VQRDMLSH: { /* VFMA, VFMS: fused multiply-add */ -- 2.20.1 Convert the Neon floating point VFMA and VFMS insn to decodetree. These are the last insns in the 3-reg-same group so we can remove all the support/loop code from the old decoder. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200512163904.10918-18-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 3 + target/arm/translate-neon.inc.c | 41 ++++++++ target/arm/translate.c | 176 +------------------------------- 3 files changed, 46 insertions(+), 174 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 @@ SHA256H2_3s 1111 001 1 0 . 01 .... .... 1100 . 1 . 0 .... \ SHA256SU1_3s 1111 001 1 0 . 10 .... .... 1100 . 1 . 0 .... \ vm=%vm_dp vn=%vn_dp vd=%vd_dp +VFMA_fp_3s 1111 001 0 0 . 0 . .... .... 1100 ... 1 .... @3same_fp +VFMS_fp_3s 1111 001 0 0 . 1 . .... .... 1100 ... 1 .... @3same_fp + VQRDMLSH_3s 1111 001 1 0 . .. .... .... 1100 ... 1 .... @3same VADD_fp_3s 1111 001 0 0 . 0 . .... .... 1101 ... 0 .... @3same_fp 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_VRSQRTS_fp_3s(DisasContext *s, arg_3same *a) return do_3same(s, a, gen_VRSQRTS_fp_3s); } +static void gen_VFMA_fp_3s(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, + TCGv_ptr fpstatus) +{ + gen_helper_vfp_muladds(vd, vn, vm, vd, fpstatus); +} + +static bool trans_VFMA_fp_3s(DisasContext *s, arg_3same *a) +{ + if (!dc_isar_feature(aa32_simdfmac, s)) { + return false; + } + + if (a->size != 0) { + /* TODO fp16 support */ + return false; + } + + return do_3same_fp(s, a, gen_VFMA_fp_3s, true); +} + +static void gen_VFMS_fp_3s(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, + TCGv_ptr fpstatus) +{ + gen_helper_vfp_negs(vn, vn); + gen_helper_vfp_muladds(vd, vn, vm, vd, fpstatus); +} + +static bool trans_VFMS_fp_3s(DisasContext *s, arg_3same *a) +{ + if (!dc_isar_feature(aa32_simdfmac, s)) { + return false; + } + + if (a->size != 0) { + /* TODO fp16 support */ + return false; + } + + return do_3same_fp(s, a, gen_VFMS_fp_3s, true); +} + static bool do_3same_fp_pair(DisasContext *s, arg_3same *a, VFPGen3OpSPFn *fn) { /* FP operations handled pairwise 32 bits at a time */ 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_narrow_op(int op, int u, int size, } } -/* Symbolic constants for op fields for Neon 3-register same-length. - * The values correspond to bits [11:8,4]; see the ARM ARM DDI0406B - * table A7-9. - */ -#define NEON_3R_VHADD 0 -#define NEON_3R_VQADD 1 -#define NEON_3R_VRHADD 2 -#define NEON_3R_LOGIC 3 /* VAND,VBIC,VORR,VMOV,VORN,VEOR,VBIF,VBIT,VBSL */ -#define NEON_3R_VHSUB 4 -#define NEON_3R_VQSUB 5 -#define NEON_3R_VCGT 6 -#define NEON_3R_VCGE 7 -#define NEON_3R_VSHL 8 -#define NEON_3R_VQSHL 9 -#define NEON_3R_VRSHL 10 -#define NEON_3R_VQRSHL 11 -#define NEON_3R_VMAX 12 -#define NEON_3R_VMIN 13 -#define NEON_3R_VABD 14 -#define NEON_3R_VABA 15 -#define NEON_3R_VADD_VSUB 16 -#define NEON_3R_VTST_VCEQ 17 -#define NEON_3R_VML 18 /* VMLA, VMLS */ -#define NEON_3R_VMUL 19 -#define NEON_3R_VPMAX 20 -#define NEON_3R_VPMIN 21 -#define NEON_3R_VQDMULH_VQRDMULH 22 -#define NEON_3R_VPADD_VQRDMLAH 23 -#define NEON_3R_SHA 24 /* SHA1C,SHA1P,SHA1M,SHA1SU0,SHA256H{2},SHA256SU1 */ -#define NEON_3R_VFM_VQRDMLSH 25 /* VFMA, VFMS, VQRDMLSH */ -#define NEON_3R_FLOAT_ARITH 26 /* float VADD, VSUB, VPADD, VABD */ -#define NEON_3R_FLOAT_MULTIPLY 27 /* float VMLA, VMLS, VMUL */ -#define NEON_3R_FLOAT_CMP 28 /* float VCEQ, VCGE, VCGT */ -#define NEON_3R_FLOAT_ACMP 29 /* float VACGE, VACGT, VACLE, VACLT */ -#define NEON_3R_FLOAT_MINMAX 30 /* float VMIN, VMAX */ -#define NEON_3R_FLOAT_MISC 31 /* float VRECPS, VRSQRTS, VMAXNM/MINNM */ - -static const uint8_t neon_3r_sizes[] = { - [NEON_3R_VHADD] = 0x7, - [NEON_3R_VQADD] = 0xf, - [NEON_3R_VRHADD] = 0x7, - [NEON_3R_LOGIC] = 0xf, /* size field encodes op type */ - [NEON_3R_VHSUB] = 0x7, - [NEON_3R_VQSUB] = 0xf, - [NEON_3R_VCGT] = 0x7, - [NEON_3R_VCGE] = 0x7, - [NEON_3R_VSHL] = 0xf, - [NEON_3R_VQSHL] = 0xf, - [NEON_3R_VRSHL] = 0xf, - [NEON_3R_VQRSHL] = 0xf, - [NEON_3R_VMAX] = 0x7, - [NEON_3R_VMIN] = 0x7, - [NEON_3R_VABD] = 0x7, - [NEON_3R_VABA] = 0x7, - [NEON_3R_VADD_VSUB] = 0xf, - [NEON_3R_VTST_VCEQ] = 0x7, - [NEON_3R_VML] = 0x7, - [NEON_3R_VMUL] = 0x7, - [NEON_3R_VPMAX] = 0x7, - [NEON_3R_VPMIN] = 0x7, - [NEON_3R_VQDMULH_VQRDMULH] = 0x6, - [NEON_3R_VPADD_VQRDMLAH] = 0x7, - [NEON_3R_SHA] = 0xf, /* size field encodes op type */ - [NEON_3R_VFM_VQRDMLSH] = 0x7, /* For VFM, size bit 1 encodes op */ - [NEON_3R_FLOAT_ARITH] = 0x5, /* size bit 1 encodes op */ - [NEON_3R_FLOAT_MULTIPLY] = 0x5, /* size bit 1 encodes op */ - [NEON_3R_FLOAT_CMP] = 0x5, /* size bit 1 encodes op */ - [NEON_3R_FLOAT_ACMP] = 0x5, /* size bit 1 encodes op */ - [NEON_3R_FLOAT_MINMAX] = 0x5, /* size bit 1 encodes op */ - [NEON_3R_FLOAT_MISC] = 0x5, /* size bit 1 encodes op */ -}; - /* Symbolic constants for op fields for Neon 2-register miscellaneous. * The values correspond to bits [17:16,10:7]; see the ARM ARM DDI0406B * table A7-13. @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) rm_ofs = neon_reg_offset(rm, 0); if ((insn & (1 << 23)) == 0) { - /* Three register same length. */ - op = ((insn >> 7) & 0x1e) | ((insn >> 4) & 1); - /* Catch invalid op and bad size combinations: UNDEF */ - if ((neon_3r_sizes[op] & (1 << size)) == 0) { - return 1; - } - /* All insns of this form UNDEF for either this condition or the - * superset of cases "Q==1"; we catch the latter later. - */ - if (q && ((rd | rn | rm) & 1)) { - return 1; - } - switch (op) { - case NEON_3R_VFM_VQRDMLSH: - if (!u) { - /* VFM, VFMS */ - if (size == 1) { - return 1; - } - break; - } - /* VQRDMLSH : handled by decodetree */ - return 1; - - case NEON_3R_VADD_VSUB: - 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: - case NEON_3R_VQADD: - case NEON_3R_VQSUB: - case NEON_3R_VMUL: - case NEON_3R_VML: - case NEON_3R_VSHL: - case NEON_3R_SHA: - case NEON_3R_VHADD: - case NEON_3R_VRHADD: - case NEON_3R_VHSUB: - case NEON_3R_VABD: - case NEON_3R_VABA: - case NEON_3R_VQSHL: - case NEON_3R_VRSHL: - case NEON_3R_VQRSHL: - case NEON_3R_VPMAX: - case NEON_3R_VPMIN: - case NEON_3R_VPADD_VQRDMLAH: - case NEON_3R_VQDMULH_VQRDMULH: - case NEON_3R_FLOAT_ARITH: - case NEON_3R_FLOAT_MULTIPLY: - case NEON_3R_FLOAT_CMP: - case NEON_3R_FLOAT_ACMP: - case NEON_3R_FLOAT_MINMAX: - case NEON_3R_FLOAT_MISC: - /* Already handled by decodetree */ - return 1; - } - - if (size == 3) { - /* 64-bit element instructions: handled by decodetree */ - return 1; - } - switch (op) { - case NEON_3R_VFM_VQRDMLSH: - if (!dc_isar_feature(aa32_simdfmac, s)) { - return 1; - } - break; - default: - break; - } - - for (pass = 0; pass < (q ? 4 : 2); pass++) { - - /* Elementwise. */ - tmp = neon_load_reg(rn, pass); - tmp2 = neon_load_reg(rm, pass); - switch (op) { - case NEON_3R_VFM_VQRDMLSH: - { - /* VFMA, VFMS: fused multiply-add */ - TCGv_ptr fpstatus = get_fpstatus_ptr(1); - TCGv_i32 tmp3 = neon_load_reg(rd, pass); - if (size) { - /* VFMS */ - gen_helper_vfp_negs(tmp, tmp); - } - gen_helper_vfp_muladds(tmp, tmp, tmp2, tmp3, fpstatus); - tcg_temp_free_i32(tmp3); - tcg_temp_free_ptr(fpstatus); - break; - } - default: - abort(); - } - tcg_temp_free_i32(tmp2); - - neon_store_reg(rd, pass, tmp); - - } /* for pass */ - /* End of 3 register same size operations. */ + /* Three register same length: handled by decodetree */ + return 1; } else if (insn & (1 << 4)) { if ((insn & 0x00380080) != 0) { /* Two registers and shift. */ -- 2.20.1 The following changes since commit 8f6330a807f2642dc2a3cdf33347aa28a4c00a87: Merge tag 'pull-maintainer-updates-060324-1' of https://gitlab.com/stsquad/qemu into staging (2024-03-06 16:56:20 +0000) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20240308 for you to fetch changes up to bbf6c6dbead82292a20951eb1204442a6b838de9: target/arm: Move v7m-related code from cpu32.c into a separate file (2024-03-08 14:45:03 +0000) ---------------------------------------------------------------- target-arm queue: * Implement FEAT_ECV * STM32L4x5: Implement GPIO device * Fix 32-bit SMOPA * Refactor v7m related code from cpu32.c into its own file * hw/rtc/sun4v-rtc: Relicense to GPLv2-or-later ---------------------------------------------------------------- Inès Varhol (3): hw/gpio: Implement STM32L4x5 GPIO hw/arm: Connect STM32L4x5 GPIO to STM32L4x5 SoC tests/qtest: Add STM32L4x5 GPIO QTest testcase Peter Maydell (9): target/arm: Move some register related defines to internals.h target/arm: Timer _EL02 registers UNDEF for E2H == 0 target/arm: use FIELD macro for CNTHCTL bit definitions target/arm: Don't allow RES0 CNTHCTL_EL2 bits to be written target/arm: Implement new FEAT_ECV trap bits target/arm: Define CNTPCTSS_EL0 and CNTVCTSS_EL0 target/arm: Implement FEAT_ECV CNTPOFF_EL2 handling target/arm: Enable FEAT_ECV for 'max' CPU hw/rtc/sun4v-rtc: Relicense to GPLv2-or-later Richard Henderson (1): target/arm: Fix 32-bit SMOPA Thomas Huth (1): target/arm: Move v7m-related code from cpu32.c into a separate file MAINTAINERS | 1 + docs/system/arm/b-l475e-iot01a.rst | 2 +- docs/system/arm/emulation.rst | 1 + include/hw/arm/stm32l4x5_soc.h | 2 + include/hw/gpio/stm32l4x5_gpio.h | 71 +++++ include/hw/misc/stm32l4x5_syscfg.h | 3 +- include/hw/rtc/sun4v-rtc.h | 2 +- target/arm/cpu-features.h | 10 + target/arm/cpu.h | 129 +-------- target/arm/internals.h | 151 ++++++++++ hw/arm/stm32l4x5_soc.c | 71 ++++- hw/gpio/stm32l4x5_gpio.c | 477 ++++++++++++++++++++++++++++++++ hw/misc/stm32l4x5_syscfg.c | 1 + hw/rtc/sun4v-rtc.c | 2 +- target/arm/helper.c | 189 ++++++++++++- target/arm/tcg/cpu-v7m.c | 290 +++++++++++++++++++ target/arm/tcg/cpu32.c | 261 ------------------ target/arm/tcg/cpu64.c | 1 + target/arm/tcg/sme_helper.c | 77 +++--- tests/qtest/stm32l4x5_gpio-test.c | 551 +++++++++++++++++++++++++++++++++++++ tests/tcg/aarch64/sme-smopa-1.c | 47 ++++ tests/tcg/aarch64/sme-smopa-2.c | 54 ++++ hw/arm/Kconfig | 3 +- hw/gpio/Kconfig | 3 + hw/gpio/meson.build | 1 + hw/gpio/trace-events | 6 + target/arm/meson.build | 3 + target/arm/tcg/meson.build | 3 + target/arm/trace-events | 1 + tests/qtest/meson.build | 3 +- tests/tcg/aarch64/Makefile.target | 2 +- 31 files changed, 1962 insertions(+), 456 deletions(-) create mode 100644 include/hw/gpio/stm32l4x5_gpio.h create mode 100644 hw/gpio/stm32l4x5_gpio.c create mode 100644 target/arm/tcg/cpu-v7m.c create mode 100644 tests/qtest/stm32l4x5_gpio-test.c create mode 100644 tests/tcg/aarch64/sme-smopa-1.c create mode 100644 tests/tcg/aarch64/sme-smopa-2.c cpu.h has a lot of #defines relating to CPU register fields. Most of these aren't actually used outside target/arm code, so there's no point in cluttering up the cpu.h file with them. Move some easy ones to internals.h. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20240301183219.2424889-2-peter.maydell@linaro.org --- target/arm/cpu.h | 128 ----------------------------------------- target/arm/internals.h | 128 +++++++++++++++++++++++++++++++++++++++++ 2 files changed, 128 insertions(+), 128 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 struct ARMGenericTimer { uint64_t ctl; /* Timer Control register */ } ARMGenericTimer; -#define VTCR_NSW (1u << 29) -#define VTCR_NSA (1u << 30) -#define VSTCR_SW VTCR_NSW -#define VSTCR_SA VTCR_NSA - /* Define a maximum sized vector register. * For 32-bit, this is a 128-bit NEON/AdvSIMD register. * For 64-bit, this is a 2048-bit SVE register. @@ -XXX,XX +XXX,XX @@ void pmu_init(ARMCPU *cpu); #define SCTLR_SPINTMASK (1ULL << 62) /* FEAT_NMI */ #define SCTLR_TIDCP (1ULL << 63) /* FEAT_TIDCP1 */ -/* Bit definitions for CPACR (AArch32 only) */ -FIELD(CPACR, CP10, 20, 2) -FIELD(CPACR, CP11, 22, 2) -FIELD(CPACR, TRCDIS, 28, 1) /* matches CPACR_EL1.TTA */ -FIELD(CPACR, D32DIS, 30, 1) /* up to v7; RAZ in v8 */ -FIELD(CPACR, ASEDIS, 31, 1) - -/* Bit definitions for CPACR_EL1 (AArch64 only) */ -FIELD(CPACR_EL1, ZEN, 16, 2) -FIELD(CPACR_EL1, FPEN, 20, 2) -FIELD(CPACR_EL1, SMEN, 24, 2) -FIELD(CPACR_EL1, TTA, 28, 1) /* matches CPACR.TRCDIS */ - -/* Bit definitions for HCPTR (AArch32 only) */ -FIELD(HCPTR, TCP10, 10, 1) -FIELD(HCPTR, TCP11, 11, 1) -FIELD(HCPTR, TASE, 15, 1) -FIELD(HCPTR, TTA, 20, 1) -FIELD(HCPTR, TAM, 30, 1) /* matches CPTR_EL2.TAM */ -FIELD(HCPTR, TCPAC, 31, 1) /* matches CPTR_EL2.TCPAC */ - -/* Bit definitions for CPTR_EL2 (AArch64 only) */ -FIELD(CPTR_EL2, TZ, 8, 1) /* !E2H */ -FIELD(CPTR_EL2, TFP, 10, 1) /* !E2H, matches HCPTR.TCP10 */ -FIELD(CPTR_EL2, TSM, 12, 1) /* !E2H */ -FIELD(CPTR_EL2, ZEN, 16, 2) /* E2H */ -FIELD(CPTR_EL2, FPEN, 20, 2) /* E2H */ -FIELD(CPTR_EL2, SMEN, 24, 2) /* E2H */ -FIELD(CPTR_EL2, TTA, 28, 1) -FIELD(CPTR_EL2, TAM, 30, 1) /* matches HCPTR.TAM */ -FIELD(CPTR_EL2, TCPAC, 31, 1) /* matches HCPTR.TCPAC */ - -/* Bit definitions for CPTR_EL3 (AArch64 only) */ -FIELD(CPTR_EL3, EZ, 8, 1) -FIELD(CPTR_EL3, TFP, 10, 1) -FIELD(CPTR_EL3, ESM, 12, 1) -FIELD(CPTR_EL3, TTA, 20, 1) -FIELD(CPTR_EL3, TAM, 30, 1) -FIELD(CPTR_EL3, TCPAC, 31, 1) - -#define MDCR_MTPME (1U << 28) -#define MDCR_TDCC (1U << 27) -#define MDCR_HLP (1U << 26) /* MDCR_EL2 */ -#define MDCR_SCCD (1U << 23) /* MDCR_EL3 */ -#define MDCR_HCCD (1U << 23) /* MDCR_EL2 */ -#define MDCR_EPMAD (1U << 21) -#define MDCR_EDAD (1U << 20) -#define MDCR_TTRF (1U << 19) -#define MDCR_STE (1U << 18) /* MDCR_EL3 */ -#define MDCR_SPME (1U << 17) /* MDCR_EL3 */ -#define MDCR_HPMD (1U << 17) /* MDCR_EL2 */ -#define MDCR_SDD (1U << 16) -#define MDCR_SPD (3U << 14) -#define MDCR_TDRA (1U << 11) -#define MDCR_TDOSA (1U << 10) -#define MDCR_TDA (1U << 9) -#define MDCR_TDE (1U << 8) -#define MDCR_HPME (1U << 7) -#define MDCR_TPM (1U << 6) -#define MDCR_TPMCR (1U << 5) -#define MDCR_HPMN (0x1fU) - -/* Not all of the MDCR_EL3 bits are present in the 32-bit SDCR */ -#define SDCR_VALID_MASK (MDCR_MTPME | MDCR_TDCC | MDCR_SCCD | \ - MDCR_EPMAD | MDCR_EDAD | MDCR_TTRF | \ - MDCR_STE | MDCR_SPME | MDCR_SPD) - #define CPSR_M (0x1fU) #define CPSR_T (1U << 5) #define CPSR_F (1U << 6) @@ -XXX,XX +XXX,XX @@ FIELD(CPTR_EL3, TCPAC, 31, 1) #define XPSR_NZCV CPSR_NZCV #define XPSR_IT CPSR_IT -#define TTBCR_N (7U << 0) /* TTBCR.EAE==0 */ -#define TTBCR_T0SZ (7U << 0) /* TTBCR.EAE==1 */ -#define TTBCR_PD0 (1U << 4) -#define TTBCR_PD1 (1U << 5) -#define TTBCR_EPD0 (1U << 7) -#define TTBCR_IRGN0 (3U << 8) -#define TTBCR_ORGN0 (3U << 10) -#define TTBCR_SH0 (3U << 12) -#define TTBCR_T1SZ (3U << 16) -#define TTBCR_A1 (1U << 22) -#define TTBCR_EPD1 (1U << 23) -#define TTBCR_IRGN1 (3U << 24) -#define TTBCR_ORGN1 (3U << 26) -#define TTBCR_SH1 (1U << 28) -#define TTBCR_EAE (1U << 31) - -FIELD(VTCR, T0SZ, 0, 6) -FIELD(VTCR, SL0, 6, 2) -FIELD(VTCR, IRGN0, 8, 2) -FIELD(VTCR, ORGN0, 10, 2) -FIELD(VTCR, SH0, 12, 2) -FIELD(VTCR, TG0, 14, 2) -FIELD(VTCR, PS, 16, 3) -FIELD(VTCR, VS, 19, 1) -FIELD(VTCR, HA, 21, 1) -FIELD(VTCR, HD, 22, 1) -FIELD(VTCR, HWU59, 25, 1) -FIELD(VTCR, HWU60, 26, 1) -FIELD(VTCR, HWU61, 27, 1) -FIELD(VTCR, HWU62, 28, 1) -FIELD(VTCR, NSW, 29, 1) -FIELD(VTCR, NSA, 30, 1) -FIELD(VTCR, DS, 32, 1) -FIELD(VTCR, SL2, 33, 1) - /* Bit definitions for ARMv8 SPSR (PSTATE) format. * Only these are valid when in AArch64 mode; in * AArch32 mode SPSRs are basically CPSR-format. @@ -XXX,XX +XXX,XX @@ static inline void xpsr_write(CPUARMState *env, uint32_t val, uint32_t mask) #define HCR_TWEDEN (1ULL << 59) #define HCR_TWEDEL MAKE_64BIT_MASK(60, 4) -#define HCRX_ENAS0 (1ULL << 0) -#define HCRX_ENALS (1ULL << 1) -#define HCRX_ENASR (1ULL << 2) -#define HCRX_FNXS (1ULL << 3) -#define HCRX_FGTNXS (1ULL << 4) -#define HCRX_SMPME (1ULL << 5) -#define HCRX_TALLINT (1ULL << 6) -#define HCRX_VINMI (1ULL << 7) -#define HCRX_VFNMI (1ULL << 8) -#define HCRX_CMOW (1ULL << 9) -#define HCRX_MCE2 (1ULL << 10) -#define HCRX_MSCEN (1ULL << 11) - -#define HPFAR_NS (1ULL << 63) - #define SCR_NS (1ULL << 0) #define SCR_IRQ (1ULL << 1) #define SCR_FIQ (1ULL << 2) @@ -XXX,XX +XXX,XX @@ static inline void xpsr_write(CPUARMState *env, uint32_t val, uint32_t mask) #define SCR_GPF (1ULL << 48) #define SCR_NSE (1ULL << 62) -#define HSTR_TTEE (1 << 16) -#define HSTR_TJDBX (1 << 17) - -#define CNTHCTL_CNTVMASK (1 << 18) -#define CNTHCTL_CNTPMASK (1 << 19) - /* Return the current FPSCR value. */ uint32_t vfp_get_fpscr(CPUARMState *env); void vfp_set_fpscr(CPUARMState *env, uint32_t val); 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(DBGWCR, WT, 20, 1) FIELD(DBGWCR, MASK, 24, 5) FIELD(DBGWCR, SSCE, 29, 1) +#define VTCR_NSW (1u << 29) +#define VTCR_NSA (1u << 30) +#define VSTCR_SW VTCR_NSW +#define VSTCR_SA VTCR_NSA + +/* Bit definitions for CPACR (AArch32 only) */ +FIELD(CPACR, CP10, 20, 2) +FIELD(CPACR, CP11, 22, 2) +FIELD(CPACR, TRCDIS, 28, 1) /* matches CPACR_EL1.TTA */ +FIELD(CPACR, D32DIS, 30, 1) /* up to v7; RAZ in v8 */ +FIELD(CPACR, ASEDIS, 31, 1) + +/* Bit definitions for CPACR_EL1 (AArch64 only) */ +FIELD(CPACR_EL1, ZEN, 16, 2) +FIELD(CPACR_EL1, FPEN, 20, 2) +FIELD(CPACR_EL1, SMEN, 24, 2) +FIELD(CPACR_EL1, TTA, 28, 1) /* matches CPACR.TRCDIS */ + +/* Bit definitions for HCPTR (AArch32 only) */ +FIELD(HCPTR, TCP10, 10, 1) +FIELD(HCPTR, TCP11, 11, 1) +FIELD(HCPTR, TASE, 15, 1) +FIELD(HCPTR, TTA, 20, 1) +FIELD(HCPTR, TAM, 30, 1) /* matches CPTR_EL2.TAM */ +FIELD(HCPTR, TCPAC, 31, 1) /* matches CPTR_EL2.TCPAC */ + +/* Bit definitions for CPTR_EL2 (AArch64 only) */ +FIELD(CPTR_EL2, TZ, 8, 1) /* !E2H */ +FIELD(CPTR_EL2, TFP, 10, 1) /* !E2H, matches HCPTR.TCP10 */ +FIELD(CPTR_EL2, TSM, 12, 1) /* !E2H */ +FIELD(CPTR_EL2, ZEN, 16, 2) /* E2H */ +FIELD(CPTR_EL2, FPEN, 20, 2) /* E2H */ +FIELD(CPTR_EL2, SMEN, 24, 2) /* E2H */ +FIELD(CPTR_EL2, TTA, 28, 1) +FIELD(CPTR_EL2, TAM, 30, 1) /* matches HCPTR.TAM */ +FIELD(CPTR_EL2, TCPAC, 31, 1) /* matches HCPTR.TCPAC */ + +/* Bit definitions for CPTR_EL3 (AArch64 only) */ +FIELD(CPTR_EL3, EZ, 8, 1) +FIELD(CPTR_EL3, TFP, 10, 1) +FIELD(CPTR_EL3, ESM, 12, 1) +FIELD(CPTR_EL3, TTA, 20, 1) +FIELD(CPTR_EL3, TAM, 30, 1) +FIELD(CPTR_EL3, TCPAC, 31, 1) + +#define MDCR_MTPME (1U << 28) +#define MDCR_TDCC (1U << 27) +#define MDCR_HLP (1U << 26) /* MDCR_EL2 */ +#define MDCR_SCCD (1U << 23) /* MDCR_EL3 */ +#define MDCR_HCCD (1U << 23) /* MDCR_EL2 */ +#define MDCR_EPMAD (1U << 21) +#define MDCR_EDAD (1U << 20) +#define MDCR_TTRF (1U << 19) +#define MDCR_STE (1U << 18) /* MDCR_EL3 */ +#define MDCR_SPME (1U << 17) /* MDCR_EL3 */ +#define MDCR_HPMD (1U << 17) /* MDCR_EL2 */ +#define MDCR_SDD (1U << 16) +#define MDCR_SPD (3U << 14) +#define MDCR_TDRA (1U << 11) +#define MDCR_TDOSA (1U << 10) +#define MDCR_TDA (1U << 9) +#define MDCR_TDE (1U << 8) +#define MDCR_HPME (1U << 7) +#define MDCR_TPM (1U << 6) +#define MDCR_TPMCR (1U << 5) +#define MDCR_HPMN (0x1fU) + +/* Not all of the MDCR_EL3 bits are present in the 32-bit SDCR */ +#define SDCR_VALID_MASK (MDCR_MTPME | MDCR_TDCC | MDCR_SCCD | \ + MDCR_EPMAD | MDCR_EDAD | MDCR_TTRF | \ + MDCR_STE | MDCR_SPME | MDCR_SPD) + +#define TTBCR_N (7U << 0) /* TTBCR.EAE==0 */ +#define TTBCR_T0SZ (7U << 0) /* TTBCR.EAE==1 */ +#define TTBCR_PD0 (1U << 4) +#define TTBCR_PD1 (1U << 5) +#define TTBCR_EPD0 (1U << 7) +#define TTBCR_IRGN0 (3U << 8) +#define TTBCR_ORGN0 (3U << 10) +#define TTBCR_SH0 (3U << 12) +#define TTBCR_T1SZ (3U << 16) +#define TTBCR_A1 (1U << 22) +#define TTBCR_EPD1 (1U << 23) +#define TTBCR_IRGN1 (3U << 24) +#define TTBCR_ORGN1 (3U << 26) +#define TTBCR_SH1 (1U << 28) +#define TTBCR_EAE (1U << 31) + +FIELD(VTCR, T0SZ, 0, 6) +FIELD(VTCR, SL0, 6, 2) +FIELD(VTCR, IRGN0, 8, 2) +FIELD(VTCR, ORGN0, 10, 2) +FIELD(VTCR, SH0, 12, 2) +FIELD(VTCR, TG0, 14, 2) +FIELD(VTCR, PS, 16, 3) +FIELD(VTCR, VS, 19, 1) +FIELD(VTCR, HA, 21, 1) +FIELD(VTCR, HD, 22, 1) +FIELD(VTCR, HWU59, 25, 1) +FIELD(VTCR, HWU60, 26, 1) +FIELD(VTCR, HWU61, 27, 1) +FIELD(VTCR, HWU62, 28, 1) +FIELD(VTCR, NSW, 29, 1) +FIELD(VTCR, NSA, 30, 1) +FIELD(VTCR, DS, 32, 1) +FIELD(VTCR, SL2, 33, 1) + +#define HCRX_ENAS0 (1ULL << 0) +#define HCRX_ENALS (1ULL << 1) +#define HCRX_ENASR (1ULL << 2) +#define HCRX_FNXS (1ULL << 3) +#define HCRX_FGTNXS (1ULL << 4) +#define HCRX_SMPME (1ULL << 5) +#define HCRX_TALLINT (1ULL << 6) +#define HCRX_VINMI (1ULL << 7) +#define HCRX_VFNMI (1ULL << 8) +#define HCRX_CMOW (1ULL << 9) +#define HCRX_MCE2 (1ULL << 10) +#define HCRX_MSCEN (1ULL << 11) + +#define HPFAR_NS (1ULL << 63) + +#define HSTR_TTEE (1 << 16) +#define HSTR_TJDBX (1 << 17) + +#define CNTHCTL_CNTVMASK (1 << 18) +#define CNTHCTL_CNTPMASK (1 << 19) + /* We use a few fake FSR values for internal purposes in M profile. * M profile cores don't have A/R format FSRs, but currently our * get_phys_addr() code assumes A/R profile and reports failures via -- 2.34.1 The timer _EL02 registers should UNDEF for invalid accesses from EL2 or EL3 when HCR_EL2.E2H == 0, not take a cp access trap. We were delivering the exception to EL2 with the wrong syndrome. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20240301183219.2424889-3-peter.maydell@linaro.org --- target/arm/helper.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static CPAccessResult e2h_access(CPUARMState *env, const ARMCPRegInfo *ri, return CP_ACCESS_OK; } if (!(arm_hcr_el2_eff(env) & HCR_E2H)) { - return CP_ACCESS_TRAP; + return CP_ACCESS_TRAP_UNCATEGORIZED; } return CP_ACCESS_OK; } -- 2.34.1 We prefer the FIELD macro over ad-hoc #defines for register bits; switch CNTHCTL to that style before we add any more bits. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20240301183219.2424889-4-peter.maydell@linaro.org --- target/arm/internals.h | 27 +++++++++++++++++++++++++-- target/arm/helper.c | 9 ++++----- 2 files changed, 29 insertions(+), 7 deletions(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ FIELD(VTCR, SL2, 33, 1) #define HSTR_TTEE (1 << 16) #define HSTR_TJDBX (1 << 17) -#define CNTHCTL_CNTVMASK (1 << 18) -#define CNTHCTL_CNTPMASK (1 << 19) +/* + * Depending on the value of HCR_EL2.E2H, bits 0 and 1 + * have different bit definitions, and EL1PCTEN might be + * bit 0 or bit 10. We use _E2H1 and _E2H0 suffixes to + * disambiguate if necessary. + */ +FIELD(CNTHCTL, EL0PCTEN_E2H1, 0, 1) +FIELD(CNTHCTL, EL0VCTEN_E2H1, 1, 1) +FIELD(CNTHCTL, EL1PCTEN_E2H0, 0, 1) +FIELD(CNTHCTL, EL1PCEN_E2H0, 1, 1) +FIELD(CNTHCTL, EVNTEN, 2, 1) +FIELD(CNTHCTL, EVNTDIR, 3, 1) +FIELD(CNTHCTL, EVNTI, 4, 4) +FIELD(CNTHCTL, EL0VTEN, 8, 1) +FIELD(CNTHCTL, EL0PTEN, 9, 1) +FIELD(CNTHCTL, EL1PCTEN_E2H1, 10, 1) +FIELD(CNTHCTL, EL1PTEN, 11, 1) +FIELD(CNTHCTL, ECV, 12, 1) +FIELD(CNTHCTL, EL1TVT, 13, 1) +FIELD(CNTHCTL, EL1TVCT, 14, 1) +FIELD(CNTHCTL, EL1NVPCT, 15, 1) +FIELD(CNTHCTL, EL1NVVCT, 16, 1) +FIELD(CNTHCTL, EVNTIS, 17, 1) +FIELD(CNTHCTL, CNTVMASK, 18, 1) +FIELD(CNTHCTL, CNTPMASK, 19, 1) /* We use a few fake FSR values for internal purposes in M profile. * M profile cores don't have A/R format FSRs, but currently our 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 gt_update_irq(ARMCPU *cpu, int timeridx) * It is RES0 in Secure and NonSecure state. */ if ((ss == ARMSS_Root || ss == ARMSS_Realm) && - ((timeridx == GTIMER_VIRT && (cnthctl & CNTHCTL_CNTVMASK)) || - (timeridx == GTIMER_PHYS && (cnthctl & CNTHCTL_CNTPMASK)))) { + ((timeridx == GTIMER_VIRT && (cnthctl & R_CNTHCTL_CNTVMASK_MASK)) || + (timeridx == GTIMER_PHYS && (cnthctl & R_CNTHCTL_CNTPMASK_MASK)))) { irqstate = 0; } @@ -XXX,XX +XXX,XX @@ static void gt_cnthctl_write(CPUARMState *env, const ARMCPRegInfo *ri, { ARMCPU *cpu = env_archcpu(env); uint32_t oldval = env->cp15.cnthctl_el2; - raw_write(env, ri, value); - if ((oldval ^ value) & CNTHCTL_CNTVMASK) { + if ((oldval ^ value) & R_CNTHCTL_CNTVMASK_MASK) { gt_update_irq(cpu, GTIMER_VIRT); - } else if ((oldval ^ value) & CNTHCTL_CNTPMASK) { + } else if ((oldval ^ value) & R_CNTHCTL_CNTPMASK_MASK) { gt_update_irq(cpu, GTIMER_PHYS); } } -- 2.34.1 Don't allow the guest to write CNTHCTL_EL2 bits which don't exist. This is not strictly architecturally required, but it is how we've tended to implement registers more recently. In particular, bits [19:18] are only present with FEAT_RME, and bits [17:12] will only be present with FEAT_ECV. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20240301183219.2424889-5-peter.maydell@linaro.org --- target/arm/helper.c | 18 ++++++++++++++++++ 1 file changed, 18 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 gt_cnthctl_write(CPUARMState *env, const ARMCPRegInfo *ri, { ARMCPU *cpu = env_archcpu(env); uint32_t oldval = env->cp15.cnthctl_el2; + uint32_t valid_mask = + R_CNTHCTL_EL0PCTEN_E2H1_MASK | + R_CNTHCTL_EL0VCTEN_E2H1_MASK | + R_CNTHCTL_EVNTEN_MASK | + R_CNTHCTL_EVNTDIR_MASK | + R_CNTHCTL_EVNTI_MASK | + R_CNTHCTL_EL0VTEN_MASK | + R_CNTHCTL_EL0PTEN_MASK | + R_CNTHCTL_EL1PCTEN_E2H1_MASK | + R_CNTHCTL_EL1PTEN_MASK; + + if (cpu_isar_feature(aa64_rme, cpu)) { + valid_mask |= R_CNTHCTL_CNTVMASK_MASK | R_CNTHCTL_CNTPMASK_MASK; + } + + /* Clear RES0 bits */ + value &= valid_mask; + raw_write(env, ri, value); if ((oldval ^ value) & R_CNTHCTL_CNTVMASK_MASK) { -- 2.34.1 The functionality defined by ID_AA64MMFR0_EL1.ECV == 1 is: * four new trap bits for various counter and timer registers * the CNTHCTL_EL2.EVNTIS and CNTKCTL_EL1.EVNTIS bits which control scaling of the event stream. This is a no-op for us, because we don't implement the event stream (our WFE is a NOP): all we need to do is allow CNTHCTL_EL2.ENVTIS to be read and written. * extensions to PMSCR_EL1.PCT, PMSCR_EL2.PCT, TRFCR_EL1.TS and TRFCR_EL2.TS: these are all no-ops for us, because we don't implement FEAT_SPE or FEAT_TRF. * new registers CNTPCTSS_EL0 and NCTVCTSS_EL0 which are "self-sychronizing" views of the CNTPCT_EL0 and CNTVCT_EL0, meaning that no barriers are needed around their accesses. For us these are just the same as the normal views, because all our sysregs are inherently self-sychronizing. In this commit we implement the trap handling and permit the new CNTHCTL_EL2 bits to be written. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20240301183219.2424889-6-peter.maydell@linaro.org --- target/arm/cpu-features.h | 5 ++++ target/arm/helper.c | 51 +++++++++++++++++++++++++++++++++++---- 2 files changed, 51 insertions(+), 5 deletions(-) diff --git a/target/arm/cpu-features.h b/target/arm/cpu-features.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu-features.h +++ b/target/arm/cpu-features.h @@ -XXX,XX +XXX,XX @@ static inline bool isar_feature_aa64_fgt(const ARMISARegisters *id) return FIELD_EX64(id->id_aa64mmfr0, ID_AA64MMFR0, FGT) != 0; } +static inline bool isar_feature_aa64_ecv_traps(const ARMISARegisters *id) +{ + return FIELD_EX64(id->id_aa64mmfr0, ID_AA64MMFR0, ECV) > 0; +} + static inline bool isar_feature_aa64_vh(const ARMISARegisters *id) { return FIELD_EX64(id->id_aa64mmfr1, ID_AA64MMFR1, VH) != 0; diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static CPAccessResult gt_counter_access(CPUARMState *env, int timeridx, : !extract32(env->cp15.cnthctl_el2, 0, 1))) { return CP_ACCESS_TRAP_EL2; } + if (has_el2 && timeridx == GTIMER_VIRT) { + if (FIELD_EX64(env->cp15.cnthctl_el2, CNTHCTL, EL1TVCT)) { + return CP_ACCESS_TRAP_EL2; + } + } break; } return CP_ACCESS_OK; @@ -XXX,XX +XXX,XX @@ static CPAccessResult gt_timer_access(CPUARMState *env, int timeridx, } } } + if (has_el2 && timeridx == GTIMER_VIRT) { + if (FIELD_EX64(env->cp15.cnthctl_el2, CNTHCTL, EL1TVT)) { + return CP_ACCESS_TRAP_EL2; + } + } break; } return CP_ACCESS_OK; @@ -XXX,XX +XXX,XX @@ static void gt_cnthctl_write(CPUARMState *env, const ARMCPRegInfo *ri, if (cpu_isar_feature(aa64_rme, cpu)) { valid_mask |= R_CNTHCTL_CNTVMASK_MASK | R_CNTHCTL_CNTPMASK_MASK; } + if (cpu_isar_feature(aa64_ecv_traps, cpu)) { + valid_mask |= + R_CNTHCTL_EL1TVT_MASK | + R_CNTHCTL_EL1TVCT_MASK | + R_CNTHCTL_EL1NVPCT_MASK | + R_CNTHCTL_EL1NVVCT_MASK | + R_CNTHCTL_EVNTIS_MASK; + } /* Clear RES0 bits */ value &= valid_mask; @@ -XXX,XX +XXX,XX @@ static CPAccessResult e2h_access(CPUARMState *env, const ARMCPRegInfo *ri, { if (arm_current_el(env) == 1) { /* This must be a FEAT_NV access */ - /* TODO: FEAT_ECV will need to check CNTHCTL_EL2 here */ return CP_ACCESS_OK; } if (!(arm_hcr_el2_eff(env) & HCR_E2H)) { @@ -XXX,XX +XXX,XX @@ static CPAccessResult e2h_access(CPUARMState *env, const ARMCPRegInfo *ri, return CP_ACCESS_OK; } +static CPAccessResult access_el1nvpct(CPUARMState *env, const ARMCPRegInfo *ri, + bool isread) +{ + if (arm_current_el(env) == 1) { + /* This must be a FEAT_NV access with NVx == 101 */ + if (FIELD_EX64(env->cp15.cnthctl_el2, CNTHCTL, EL1NVPCT)) { + return CP_ACCESS_TRAP_EL2; + } + } + return e2h_access(env, ri, isread); +} + +static CPAccessResult access_el1nvvct(CPUARMState *env, const ARMCPRegInfo *ri, + bool isread) +{ + if (arm_current_el(env) == 1) { + /* This must be a FEAT_NV access with NVx == 101 */ + if (FIELD_EX64(env->cp15.cnthctl_el2, CNTHCTL, EL1NVVCT)) { + return CP_ACCESS_TRAP_EL2; + } + } + return e2h_access(env, ri, isread); +} + /* Test if system register redirection is to occur in the current state. */ static bool redirect_for_e2h(CPUARMState *env) { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo vhe_reginfo[] = { { .name = "CNTP_CTL_EL02", .state = ARM_CP_STATE_AA64, .opc0 = 3, .opc1 = 5, .crn = 14, .crm = 2, .opc2 = 1, .type = ARM_CP_IO | ARM_CP_ALIAS, - .access = PL2_RW, .accessfn = e2h_access, + .access = PL2_RW, .accessfn = access_el1nvpct, .nv2_redirect_offset = 0x180 | NV2_REDIR_NO_NV1, .fieldoffset = offsetof(CPUARMState, cp15.c14_timer[GTIMER_PHYS].ctl), .writefn = gt_phys_ctl_write, .raw_writefn = raw_write }, { .name = "CNTV_CTL_EL02", .state = ARM_CP_STATE_AA64, .opc0 = 3, .opc1 = 5, .crn = 14, .crm = 3, .opc2 = 1, .type = ARM_CP_IO | ARM_CP_ALIAS, - .access = PL2_RW, .accessfn = e2h_access, + .access = PL2_RW, .accessfn = access_el1nvvct, .nv2_redirect_offset = 0x170 | NV2_REDIR_NO_NV1, .fieldoffset = offsetof(CPUARMState, cp15.c14_timer[GTIMER_VIRT].ctl), .writefn = gt_virt_ctl_write, .raw_writefn = raw_write }, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo vhe_reginfo[] = { .type = ARM_CP_IO | ARM_CP_ALIAS, .fieldoffset = offsetof(CPUARMState, cp15.c14_timer[GTIMER_PHYS].cval), .nv2_redirect_offset = 0x178 | NV2_REDIR_NO_NV1, - .access = PL2_RW, .accessfn = e2h_access, + .access = PL2_RW, .accessfn = access_el1nvpct, .writefn = gt_phys_cval_write, .raw_writefn = raw_write }, { .name = "CNTV_CVAL_EL02", .state = ARM_CP_STATE_AA64, .opc0 = 3, .opc1 = 5, .crn = 14, .crm = 3, .opc2 = 2, .type = ARM_CP_IO | ARM_CP_ALIAS, .nv2_redirect_offset = 0x168 | NV2_REDIR_NO_NV1, .fieldoffset = offsetof(CPUARMState, cp15.c14_timer[GTIMER_VIRT].cval), - .access = PL2_RW, .accessfn = e2h_access, + .access = PL2_RW, .accessfn = access_el1nvvct, .writefn = gt_virt_cval_write, .raw_writefn = raw_write }, #endif }; -- 2.34.1 For FEAT_ECV, new registers CNTPCTSS_EL0 and CNTVCTSS_EL0 are defined, which are "self-synchronized" views of the physical and virtual counts as seen in the CNTPCT_EL0 and CNTVCT_EL0 registers (meaning that no barriers are needed around accesses to them to ensure that reads of them do not occur speculatively and out-of-order with other instructions). For QEMU, all our system registers are self-synchronized, so we can simply copy the existing implementation of CNTPCT_EL0 and CNTVCT_EL0 to the new register encodings. This means we now implement all the functionality required for ID_AA64MMFR0_EL1.ECV == 0b0001. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20240301183219.2424889-7-peter.maydell@linaro.org --- target/arm/helper.c | 43 +++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 43 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 generic_timer_cp_reginfo[] = { }, }; +/* + * FEAT_ECV adds extra views of CNTVCT_EL0 and CNTPCT_EL0 which + * are "self-synchronizing". For QEMU all sysregs are self-synchronizing, + * so our implementations here are identical to the normal registers. + */ +static const ARMCPRegInfo gen_timer_ecv_cp_reginfo[] = { + { .name = "CNTVCTSS", .cp = 15, .crm = 14, .opc1 = 9, + .access = PL0_R, .type = ARM_CP_64BIT | ARM_CP_NO_RAW | ARM_CP_IO, + .accessfn = gt_vct_access, + .readfn = gt_virt_cnt_read, .resetfn = arm_cp_reset_ignore, + }, + { .name = "CNTVCTSS_EL0", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 3, .crn = 14, .crm = 0, .opc2 = 6, + .access = PL0_R, .type = ARM_CP_NO_RAW | ARM_CP_IO, + .accessfn = gt_vct_access, .readfn = gt_virt_cnt_read, + }, + { .name = "CNTPCTSS", .cp = 15, .crm = 14, .opc1 = 8, + .access = PL0_R, .type = ARM_CP_64BIT | ARM_CP_NO_RAW | ARM_CP_IO, + .accessfn = gt_pct_access, + .readfn = gt_cnt_read, .resetfn = arm_cp_reset_ignore, + }, + { .name = "CNTPCTSS_EL0", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 3, .crn = 14, .crm = 0, .opc2 = 5, + .access = PL0_R, .type = ARM_CP_NO_RAW | ARM_CP_IO, + .accessfn = gt_pct_access, .readfn = gt_cnt_read, + }, +}; + #else /* @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo generic_timer_cp_reginfo[] = { }, }; +/* + * CNTVCTSS_EL0 has the same trap conditions as CNTVCT_EL0, so it also + * is exposed to userspace by Linux. + */ +static const ARMCPRegInfo gen_timer_ecv_cp_reginfo[] = { + { .name = "CNTVCTSS_EL0", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 3, .crn = 14, .crm = 0, .opc2 = 6, + .access = PL0_R, .type = ARM_CP_NO_RAW | ARM_CP_IO, + .readfn = gt_virt_cnt_read, + }, +}; + #endif static void par_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value) @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) if (arm_feature(env, ARM_FEATURE_GENERIC_TIMER)) { define_arm_cp_regs(cpu, generic_timer_cp_reginfo); } + if (cpu_isar_feature(aa64_ecv_traps, cpu)) { + define_arm_cp_regs(cpu, gen_timer_ecv_cp_reginfo); + } if (arm_feature(env, ARM_FEATURE_VAPA)) { ARMCPRegInfo vapa_cp_reginfo[] = { { .name = "PAR", .cp = 15, .crn = 7, .crm = 4, .opc1 = 0, .opc2 = 0, -- 2.34.1 When ID_AA64MMFR0_EL1.ECV is 0b0010, a new register CNTPOFF_EL2 is implemented. This is similar to the existing CNTVOFF_EL2, except that it controls a hypervisor-adjustable offset made to the physical counter and timer. Implement the handling for this register, which includes control/trap bits in SCR_EL3 and CNTHCTL_EL2. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20240301183219.2424889-8-peter.maydell@linaro.org --- target/arm/cpu-features.h | 5 +++ target/arm/cpu.h | 1 + target/arm/helper.c | 68 +++++++++++++++++++++++++++++++++++++-- target/arm/trace-events | 1 + 4 files changed, 73 insertions(+), 2 deletions(-) diff --git a/target/arm/cpu-features.h b/target/arm/cpu-features.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu-features.h +++ b/target/arm/cpu-features.h @@ -XXX,XX +XXX,XX @@ static inline bool isar_feature_aa64_ecv_traps(const ARMISARegisters *id) return FIELD_EX64(id->id_aa64mmfr0, ID_AA64MMFR0, ECV) > 0; } +static inline bool isar_feature_aa64_ecv(const ARMISARegisters *id) +{ + return FIELD_EX64(id->id_aa64mmfr0, ID_AA64MMFR0, ECV) > 1; +} + static inline bool isar_feature_aa64_vh(const ARMISARegisters *id) { return FIELD_EX64(id->id_aa64mmfr1, ID_AA64MMFR1, VH) != 0; 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 CPUArchState { uint64_t c14_cntkctl; /* Timer Control register */ uint64_t cnthctl_el2; /* Counter/Timer Hyp Control register */ uint64_t cntvoff_el2; /* Counter Virtual Offset register */ + uint64_t cntpoff_el2; /* Counter Physical Offset register */ ARMGenericTimer c14_timer[NUM_GTIMERS]; uint32_t c15_cpar; /* XScale Coprocessor Access Register */ uint32_t c15_ticonfig; /* TI925T configuration byte. */ 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_rme, cpu)) { valid_mask |= SCR_NSE | SCR_GPF; } + if (cpu_isar_feature(aa64_ecv, cpu)) { + valid_mask |= SCR_ECVEN; + } } else { valid_mask &= ~(SCR_RW | SCR_ST); if (cpu_isar_feature(aa32_ras, cpu)) { @@ -XXX,XX +XXX,XX @@ void gt_rme_post_el_change(ARMCPU *cpu, void *ignored) gt_update_irq(cpu, GTIMER_PHYS); } +static uint64_t gt_phys_raw_cnt_offset(CPUARMState *env) +{ + if ((env->cp15.scr_el3 & SCR_ECVEN) && + FIELD_EX64(env->cp15.cnthctl_el2, CNTHCTL, ECV) && + arm_is_el2_enabled(env) && + (arm_hcr_el2_eff(env) & (HCR_E2H | HCR_TGE)) != (HCR_E2H | HCR_TGE)) { + return env->cp15.cntpoff_el2; + } + return 0; +} + +static uint64_t gt_phys_cnt_offset(CPUARMState *env) +{ + if (arm_current_el(env) >= 2) { + return 0; + } + return gt_phys_raw_cnt_offset(env); +} + static void gt_recalc_timer(ARMCPU *cpu, int timeridx) { ARMGenericTimer *gt = &cpu->env.cp15.c14_timer[timeridx]; @@ -XXX,XX +XXX,XX @@ static void gt_recalc_timer(ARMCPU *cpu, int timeridx) * reset timer to when ISTATUS next has to change */ uint64_t offset = timeridx == GTIMER_VIRT ? - cpu->env.cp15.cntvoff_el2 : 0; + cpu->env.cp15.cntvoff_el2 : gt_phys_raw_cnt_offset(&cpu->env); uint64_t count = gt_get_countervalue(&cpu->env); /* Note that this must be unsigned 64 bit arithmetic: */ int istatus = count - offset >= gt->cval; @@ -XXX,XX +XXX,XX @@ static void gt_timer_reset(CPUARMState *env, const ARMCPRegInfo *ri, static uint64_t gt_cnt_read(CPUARMState *env, const ARMCPRegInfo *ri) { - return gt_get_countervalue(env); + return gt_get_countervalue(env) - gt_phys_cnt_offset(env); } static uint64_t gt_virt_cnt_offset(CPUARMState *env) @@ -XXX,XX +XXX,XX @@ static uint64_t gt_tval_read(CPUARMState *env, const ARMCPRegInfo *ri, case GTIMER_HYPVIRT: offset = gt_virt_cnt_offset(env); break; + case GTIMER_PHYS: + offset = gt_phys_cnt_offset(env); + break; } return (uint32_t)(env->cp15.c14_timer[timeridx].cval - @@ -XXX,XX +XXX,XX @@ static void gt_tval_write(CPUARMState *env, const ARMCPRegInfo *ri, case GTIMER_HYPVIRT: offset = gt_virt_cnt_offset(env); break; + case GTIMER_PHYS: + offset = gt_phys_cnt_offset(env); + break; } trace_arm_gt_tval_write(timeridx, value); @@ -XXX,XX +XXX,XX @@ static void gt_cnthctl_write(CPUARMState *env, const ARMCPRegInfo *ri, R_CNTHCTL_EL1NVVCT_MASK | R_CNTHCTL_EVNTIS_MASK; } + if (cpu_isar_feature(aa64_ecv, cpu)) { + valid_mask |= R_CNTHCTL_ECV_MASK; + } /* Clear RES0 bits */ value &= valid_mask; @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo gen_timer_ecv_cp_reginfo[] = { }, }; +static CPAccessResult gt_cntpoff_access(CPUARMState *env, + const ARMCPRegInfo *ri, + bool isread) +{ + if (arm_current_el(env) == 2 && !(env->cp15.scr_el3 & SCR_ECVEN)) { + return CP_ACCESS_TRAP_EL3; + } + return CP_ACCESS_OK; +} + +static void gt_cntpoff_write(CPUARMState *env, const ARMCPRegInfo *ri, + uint64_t value) +{ + ARMCPU *cpu = env_archcpu(env); + + trace_arm_gt_cntpoff_write(value); + raw_write(env, ri, value); + gt_recalc_timer(cpu, GTIMER_PHYS); +} + +static const ARMCPRegInfo gen_timer_cntpoff_reginfo = { + .name = "CNTPOFF_EL2", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 4, .crn = 14, .crm = 0, .opc2 = 6, + .access = PL2_RW, .type = ARM_CP_IO, .resetvalue = 0, + .accessfn = gt_cntpoff_access, .writefn = gt_cntpoff_write, + .nv2_redirect_offset = 0x1a8, + .fieldoffset = offsetof(CPUARMState, cp15.cntpoff_el2), +}; #else /* @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) if (cpu_isar_feature(aa64_ecv_traps, cpu)) { define_arm_cp_regs(cpu, gen_timer_ecv_cp_reginfo); } +#ifndef CONFIG_USER_ONLY + if (cpu_isar_feature(aa64_ecv, cpu)) { + define_one_arm_cp_reg(cpu, &gen_timer_cntpoff_reginfo); + } +#endif if (arm_feature(env, ARM_FEATURE_VAPA)) { ARMCPRegInfo vapa_cp_reginfo[] = { { .name = "PAR", .cp = 15, .crn = 7, .crm = 4, .opc1 = 0, .opc2 = 0, diff --git a/target/arm/trace-events b/target/arm/trace-events index XXXXXXX..XXXXXXX 100644 --- a/target/arm/trace-events +++ b/target/arm/trace-events @@ -XXX,XX +XXX,XX @@ arm_gt_tval_write(int timer, uint64_t value) "gt_tval_write: timer %d value 0x%" arm_gt_ctl_write(int timer, uint64_t value) "gt_ctl_write: timer %d value 0x%" PRIx64 arm_gt_imask_toggle(int timer) "gt_ctl_write: timer %d IMASK toggle" arm_gt_cntvoff_write(uint64_t value) "gt_cntvoff_write: value 0x%" PRIx64 +arm_gt_cntpoff_write(uint64_t value) "gt_cntpoff_write: value 0x%" PRIx64 arm_gt_update_irq(int timer, int irqstate) "gt_update_irq: timer %d irqstate %d" # kvm.c -- 2.34.1 Enable all FEAT_ECV features on the 'max' CPU. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20240301183219.2424889-9-peter.maydell@linaro.org --- docs/system/arm/emulation.rst | 1 + target/arm/tcg/cpu64.c | 1 + 2 files changed, 2 insertions(+) diff --git a/docs/system/arm/emulation.rst b/docs/system/arm/emulation.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/emulation.rst +++ b/docs/system/arm/emulation.rst @@ -XXX,XX +XXX,XX @@ the following architecture extensions: - FEAT_DotProd (Advanced SIMD dot product instructions) - FEAT_DoubleFault (Double Fault Extension) - FEAT_E0PD (Preventing EL0 access to halves of address maps) +- FEAT_ECV (Enhanced Counter Virtualization) - FEAT_EPAC (Enhanced pointer authentication) - FEAT_ETS (Enhanced Translation Synchronization) - FEAT_EVT (Enhanced Virtualization Traps) diff --git a/target/arm/tcg/cpu64.c b/target/arm/tcg/cpu64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/cpu64.c +++ b/target/arm/tcg/cpu64.c @@ -XXX,XX +XXX,XX @@ void aarch64_max_tcg_initfn(Object *obj) t = FIELD_DP64(t, ID_AA64MMFR0, TGRAN64_2, 2); /* 64k stage2 supported */ t = FIELD_DP64(t, ID_AA64MMFR0, TGRAN4_2, 2); /* 4k stage2 supported */ t = FIELD_DP64(t, ID_AA64MMFR0, FGT, 1); /* FEAT_FGT */ + t = FIELD_DP64(t, ID_AA64MMFR0, ECV, 2); /* FEAT_ECV */ cpu->isar.id_aa64mmfr0 = t; t = cpu->isar.id_aa64mmfr1; -- 2.34.1 From: Inès Varhol <ines.varhol@telecom-paris.fr> Features supported : - the 8 STM32L4x5 GPIOs are initialized with their reset values (except IDR, see below) - input mode : setting a pin in input mode "externally" (using input irqs) results in an out irq (transmitted to SYSCFG) - output mode : setting a bit in ODR sets the corresponding out irq (if this line is configured in output mode) - pull-up, pull-down - push-pull, open-drain Difference with the real GPIOs : - Alternate Function and Analog mode aren't implemented : pins in AF/Analog behave like pins in input mode - floating pins stay at their last value - register IDR reset values differ from the real one : values are coherent with the other registers reset values and the fact that AF/Analog modes aren't implemented - setting I/O output speed isn't supported - locking port bits isn't supported - ADC function isn't supported - GPIOH has 16 pins instead of 2 pins - writing to registers LCKR, AFRL, AFRH and ASCR is ineffective Signed-off-by: Arnaud Minier <arnaud.minier@telecom-paris.fr> Signed-off-by: Inès Varhol <ines.varhol@telecom-paris.fr> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Acked-by: Alistair Francis <alistair.francis@wdc.com> Message-id: 20240305210444.310665-2-ines.varhol@telecom-paris.fr Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- MAINTAINERS | 1 + docs/system/arm/b-l475e-iot01a.rst | 2 +- include/hw/gpio/stm32l4x5_gpio.h | 70 +++++ hw/gpio/stm32l4x5_gpio.c | 477 +++++++++++++++++++++++++++++ hw/gpio/Kconfig | 3 + hw/gpio/meson.build | 1 + hw/gpio/trace-events | 6 + 7 files changed, 559 insertions(+), 1 deletion(-) create mode 100644 include/hw/gpio/stm32l4x5_gpio.h create mode 100644 hw/gpio/stm32l4x5_gpio.c diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ F: hw/arm/stm32l4x5_soc.c F: hw/misc/stm32l4x5_exti.c F: hw/misc/stm32l4x5_syscfg.c F: hw/misc/stm32l4x5_rcc.c +F: hw/gpio/stm32l4x5_gpio.c F: include/hw/*/stm32l4x5_*.h B-L475E-IOT01A IoT Node diff --git a/docs/system/arm/b-l475e-iot01a.rst b/docs/system/arm/b-l475e-iot01a.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/b-l475e-iot01a.rst +++ b/docs/system/arm/b-l475e-iot01a.rst @@ -XXX,XX +XXX,XX @@ Currently B-L475E-IOT01A machine's only supports the following devices: - STM32L4x5 EXTI (Extended interrupts and events controller) - STM32L4x5 SYSCFG (System configuration controller) - STM32L4x5 RCC (Reset and clock control) +- STM32L4x5 GPIOs (General-purpose I/Os) Missing devices """"""""""""""" @@ -XXX,XX +XXX,XX @@ Missing devices The B-L475E-IOT01A does *not* support the following devices: - Serial ports (UART) -- General-purpose I/Os (GPIO) - Analog to Digital Converter (ADC) - SPI controller - Timer controller (TIMER) diff --git a/include/hw/gpio/stm32l4x5_gpio.h b/include/hw/gpio/stm32l4x5_gpio.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/gpio/stm32l4x5_gpio.h @@ -XXX,XX +XXX,XX @@ +/* + * STM32L4x5 GPIO (General Purpose Input/Ouput) + * + * Copyright (c) 2024 Arnaud Minier <arnaud.minier@telecom-paris.fr> + * Copyright (c) 2024 Inès Varhol <ines.varhol@telecom-paris.fr> + * + * SPDX-License-Identifier: GPL-2.0-or-later + * + * This work is licensed under the terms of the GNU GPL, version 2 or later. + * See the COPYING file in the top-level directory. + */ + +/* + * The reference used is the STMicroElectronics RM0351 Reference manual + * for STM32L4x5 and STM32L4x6 advanced Arm ® -based 32-bit MCUs. + * https://www.st.com/en/microcontrollers-microprocessors/stm32l4x5/documentation.html + */ + +#ifndef HW_STM32L4X5_GPIO_H +#define HW_STM32L4X5_GPIO_H + +#include "hw/sysbus.h" +#include "qom/object.h" + +#define TYPE_STM32L4X5_GPIO "stm32l4x5-gpio" +OBJECT_DECLARE_SIMPLE_TYPE(Stm32l4x5GpioState, STM32L4X5_GPIO) + +#define GPIO_NUM_PINS 16 + +struct Stm32l4x5GpioState { + SysBusDevice parent_obj; + + MemoryRegion mmio; + + /* GPIO registers */ + uint32_t moder; + uint32_t otyper; + uint32_t ospeedr; + uint32_t pupdr; + uint32_t idr; + uint32_t odr; + uint32_t lckr; + uint32_t afrl; + uint32_t afrh; + uint32_t ascr; + + /* GPIO registers reset values */ + uint32_t moder_reset; + uint32_t ospeedr_reset; + uint32_t pupdr_reset; + + /* + * External driving of pins. + * The pins can be set externally through the device + * anonymous input GPIOs lines under certain conditions. + * The pin must not be in push-pull output mode, + * and can't be set high in open-drain mode. + * Pins driven externally and configured to + * output mode will in general be "disconnected" + * (see `get_gpio_pinmask_to_disconnect()`) + */ + uint16_t disconnected_pins; + uint16_t pins_connected_high; + + char *name; + Clock *clk; + qemu_irq pin[GPIO_NUM_PINS]; +}; + +#endif diff --git a/hw/gpio/stm32l4x5_gpio.c b/hw/gpio/stm32l4x5_gpio.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/gpio/stm32l4x5_gpio.c @@ -XXX,XX +XXX,XX @@ +/* + * STM32L4x5 GPIO (General Purpose Input/Ouput) + * + * Copyright (c) 2024 Arnaud Minier <arnaud.minier@telecom-paris.fr> + * Copyright (c) 2024 Inès Varhol <ines.varhol@telecom-paris.fr> + * + * SPDX-License-Identifier: GPL-2.0-or-later + * + * This work is licensed under the terms of the GNU GPL, version 2 or later. + * See the COPYING file in the top-level directory. + */ + +/* + * The reference used is the STMicroElectronics RM0351 Reference manual + * for STM32L4x5 and STM32L4x6 advanced Arm ® -based 32-bit MCUs. + * https://www.st.com/en/microcontrollers-microprocessors/stm32l4x5/documentation.html + */ + +#include "qemu/osdep.h" +#include "qemu/log.h" +#include "hw/gpio/stm32l4x5_gpio.h" +#include "hw/irq.h" +#include "hw/qdev-clock.h" +#include "hw/qdev-properties.h" +#include "qapi/visitor.h" +#include "qapi/error.h" +#include "migration/vmstate.h" +#include "trace.h" + +#define GPIO_MODER 0x00 +#define GPIO_OTYPER 0x04 +#define GPIO_OSPEEDR 0x08 +#define GPIO_PUPDR 0x0C +#define GPIO_IDR 0x10 +#define GPIO_ODR 0x14 +#define GPIO_BSRR 0x18 +#define GPIO_LCKR 0x1C +#define GPIO_AFRL 0x20 +#define GPIO_AFRH 0x24 +#define GPIO_BRR 0x28 +#define GPIO_ASCR 0x2C + +/* 0b11111111_11111111_00000000_00000000 */ +#define RESERVED_BITS_MASK 0xFFFF0000 + +static void update_gpio_idr(Stm32l4x5GpioState *s); + +static bool is_pull_up(Stm32l4x5GpioState *s, unsigned pin) +{ + return extract32(s->pupdr, 2 * pin, 2) == 1; +} + +static bool is_pull_down(Stm32l4x5GpioState *s, unsigned pin) +{ + return extract32(s->pupdr, 2 * pin, 2) == 2; +} + +static bool is_output(Stm32l4x5GpioState *s, unsigned pin) +{ + return extract32(s->moder, 2 * pin, 2) == 1; +} + +static bool is_open_drain(Stm32l4x5GpioState *s, unsigned pin) +{ + return extract32(s->otyper, pin, 1) == 1; +} + +static bool is_push_pull(Stm32l4x5GpioState *s, unsigned pin) +{ + return extract32(s->otyper, pin, 1) == 0; +} + +static void stm32l4x5_gpio_reset_hold(Object *obj) +{ + Stm32l4x5GpioState *s = STM32L4X5_GPIO(obj); + + s->moder = s->moder_reset; + s->otyper = 0x00000000; + s->ospeedr = s->ospeedr_reset; + s->pupdr = s->pupdr_reset; + s->idr = 0x00000000; + s->odr = 0x00000000; + s->lckr = 0x00000000; + s->afrl = 0x00000000; + s->afrh = 0x00000000; + s->ascr = 0x00000000; + + s->disconnected_pins = 0xFFFF; + s->pins_connected_high = 0x0000; + update_gpio_idr(s); +} + +static void stm32l4x5_gpio_set(void *opaque, int line, int level) +{ + Stm32l4x5GpioState *s = opaque; + /* + * The pin isn't set if line is configured in output mode + * except if level is 0 and the output is open-drain. + * This way there will be no short-circuit prone situations. + */ + if (is_output(s, line) && !(is_open_drain(s, line) && (level == 0))) { + qemu_log_mask(LOG_GUEST_ERROR, "Line %d can't be driven externally\n", + line); + return; + } + + s->disconnected_pins &= ~(1 << line); + if (level) { + s->pins_connected_high |= (1 << line); + } else { + s->pins_connected_high &= ~(1 << line); + } + trace_stm32l4x5_gpio_pins(s->name, s->disconnected_pins, + s->pins_connected_high); + update_gpio_idr(s); +} + + +static void update_gpio_idr(Stm32l4x5GpioState *s) +{ + uint32_t new_idr_mask = 0; + uint32_t new_idr = s->odr; + uint32_t old_idr = s->idr; + int new_pin_state, old_pin_state; + + for (int i = 0; i < GPIO_NUM_PINS; i++) { + if (is_output(s, i)) { + if (is_push_pull(s, i)) { + new_idr_mask |= (1 << i); + } else if (!(s->odr & (1 << i))) { + /* open-drain ODR 0 */ + new_idr_mask |= (1 << i); + /* open-drain ODR 1 */ + } else if (!(s->disconnected_pins & (1 << i)) && + !(s->pins_connected_high & (1 << i))) { + /* open-drain ODR 1 with pin connected low */ + new_idr_mask |= (1 << i); + new_idr &= ~(1 << i); + /* open-drain ODR 1 with unactive pin */ + } else if (is_pull_up(s, i)) { + new_idr_mask |= (1 << i); + } else if (is_pull_down(s, i)) { + new_idr_mask |= (1 << i); + new_idr &= ~(1 << i); + } + /* + * The only case left is for open-drain ODR 1 + * with unactive pin without pull-up or pull-down : + * the value is floating. + */ + /* input or analog mode with connected pin */ + } else if (!(s->disconnected_pins & (1 << i))) { + if (s->pins_connected_high & (1 << i)) { + /* pin high */ + new_idr_mask |= (1 << i); + new_idr |= (1 << i); + } else { + /* pin low */ + new_idr_mask |= (1 << i); + new_idr &= ~(1 << i); + } + /* input or analog mode with disconnected pin */ + } else { + if (is_pull_up(s, i)) { + /* pull-up */ + new_idr_mask |= (1 << i); + new_idr |= (1 << i); + } else if (is_pull_down(s, i)) { + /* pull-down */ + new_idr_mask |= (1 << i); + new_idr &= ~(1 << i); + } + /* + * The only case left is for a disconnected pin + * without pull-up or pull-down : + * the value is floating. + */ + } + } + + s->idr = (old_idr & ~new_idr_mask) | (new_idr & new_idr_mask); + trace_stm32l4x5_gpio_update_idr(s->name, old_idr, s->idr); + + for (int i = 0; i < GPIO_NUM_PINS; i++) { + if (new_idr_mask & (1 << i)) { + new_pin_state = (new_idr & (1 << i)) > 0; + old_pin_state = (old_idr & (1 << i)) > 0; + if (new_pin_state > old_pin_state) { + qemu_irq_raise(s->pin[i]); + } else if (new_pin_state < old_pin_state) { + qemu_irq_lower(s->pin[i]); + } + } + } +} + +/* + * Return mask of pins that are both configured in output + * mode and externally driven (except pins in open-drain + * mode externally set to 0). + */ +static uint32_t get_gpio_pinmask_to_disconnect(Stm32l4x5GpioState *s) +{ + uint32_t pins_to_disconnect = 0; + for (int i = 0; i < GPIO_NUM_PINS; i++) { + /* for each connected pin in output mode */ + if (!(s->disconnected_pins & (1 << i)) && is_output(s, i)) { + /* if either push-pull or high level */ + if (is_push_pull(s, i) || s->pins_connected_high & (1 << i)) { + pins_to_disconnect |= (1 << i); + qemu_log_mask(LOG_GUEST_ERROR, + "Line %d can't be driven externally\n", + i); + } + } + } + return pins_to_disconnect; +} + +/* + * Set field `disconnected_pins` and call `update_gpio_idr()` + */ +static void disconnect_gpio_pins(Stm32l4x5GpioState *s, uint16_t lines) +{ + s->disconnected_pins |= lines; + trace_stm32l4x5_gpio_pins(s->name, s->disconnected_pins, + s->pins_connected_high); + update_gpio_idr(s); +} + +static void disconnected_pins_set(Object *obj, Visitor *v, + const char *name, void *opaque, Error **errp) +{ + Stm32l4x5GpioState *s = STM32L4X5_GPIO(obj); + uint16_t value; + if (!visit_type_uint16(v, name, &value, errp)) { + return; + } + disconnect_gpio_pins(s, value); +} + +static void disconnected_pins_get(Object *obj, Visitor *v, + const char *name, void *opaque, Error **errp) +{ + visit_type_uint16(v, name, (uint16_t *)opaque, errp); +} + +static void clock_freq_get(Object *obj, Visitor *v, + const char *name, void *opaque, Error **errp) +{ + Stm32l4x5GpioState *s = STM32L4X5_GPIO(obj); + uint32_t clock_freq_hz = clock_get_hz(s->clk); + visit_type_uint32(v, name, &clock_freq_hz, errp); +} + +static void stm32l4x5_gpio_write(void *opaque, hwaddr addr, + uint64_t val64, unsigned int size) +{ + Stm32l4x5GpioState *s = opaque; + + uint32_t value = val64; + trace_stm32l4x5_gpio_write(s->name, addr, val64); + + switch (addr) { + case GPIO_MODER: + s->moder = value; + disconnect_gpio_pins(s, get_gpio_pinmask_to_disconnect(s)); + qemu_log_mask(LOG_UNIMP, + "%s: Analog and AF modes aren't supported\n\ + Analog and AF mode behave like input mode\n", + __func__); + return; + case GPIO_OTYPER: + s->otyper = value & ~RESERVED_BITS_MASK; + disconnect_gpio_pins(s, get_gpio_pinmask_to_disconnect(s)); + return; + case GPIO_OSPEEDR: + qemu_log_mask(LOG_UNIMP, + "%s: Changing I/O output speed isn't supported\n\ + I/O speed is already maximal\n", + __func__); + s->ospeedr = value; + return; + case GPIO_PUPDR: + s->pupdr = value; + update_gpio_idr(s); + return; + case GPIO_IDR: + qemu_log_mask(LOG_UNIMP, + "%s: GPIO->IDR is read-only\n", + __func__); + return; + case GPIO_ODR: + s->odr = value & ~RESERVED_BITS_MASK; + update_gpio_idr(s); + return; + case GPIO_BSRR: { + uint32_t bits_to_reset = (value & RESERVED_BITS_MASK) >> GPIO_NUM_PINS; + uint32_t bits_to_set = value & ~RESERVED_BITS_MASK; + /* If both BSx and BRx are set, BSx has priority.*/ + s->odr &= ~bits_to_reset; + s->odr |= bits_to_set; + update_gpio_idr(s); + return; + } + case GPIO_LCKR: + qemu_log_mask(LOG_UNIMP, + "%s: Locking port bits configuration isn't supported\n", + __func__); + s->lckr = value & ~RESERVED_BITS_MASK; + return; + case GPIO_AFRL: + qemu_log_mask(LOG_UNIMP, + "%s: Alternate functions aren't supported\n", + __func__); + s->afrl = value; + return; + case GPIO_AFRH: + qemu_log_mask(LOG_UNIMP, + "%s: Alternate functions aren't supported\n", + __func__); + s->afrh = value; + return; + case GPIO_BRR: { + uint32_t bits_to_reset = value & ~RESERVED_BITS_MASK; + s->odr &= ~bits_to_reset; + update_gpio_idr(s); + return; + } + case GPIO_ASCR: + qemu_log_mask(LOG_UNIMP, + "%s: ADC function isn't supported\n", + __func__); + s->ascr = value & ~RESERVED_BITS_MASK; + return; + default: + qemu_log_mask(LOG_GUEST_ERROR, + "%s: Bad offset 0x%" HWADDR_PRIx "\n", __func__, addr); + } +} + +static uint64_t stm32l4x5_gpio_read(void *opaque, hwaddr addr, + unsigned int size) +{ + Stm32l4x5GpioState *s = opaque; + + trace_stm32l4x5_gpio_read(s->name, addr); + + switch (addr) { + case GPIO_MODER: + return s->moder; + case GPIO_OTYPER: + return s->otyper; + case GPIO_OSPEEDR: + return s->ospeedr; + case GPIO_PUPDR: + return s->pupdr; + case GPIO_IDR: + return s->idr; + case GPIO_ODR: + return s->odr; + case GPIO_BSRR: + return 0; + case GPIO_LCKR: + return s->lckr; + case GPIO_AFRL: + return s->afrl; + case GPIO_AFRH: + return s->afrh; + case GPIO_BRR: + return 0; + case GPIO_ASCR: + return s->ascr; + default: + qemu_log_mask(LOG_GUEST_ERROR, + "%s: Bad offset 0x%" HWADDR_PRIx "\n", __func__, addr); + return 0; + } +} + +static const MemoryRegionOps stm32l4x5_gpio_ops = { + .read = stm32l4x5_gpio_read, + .write = stm32l4x5_gpio_write, + .endianness = DEVICE_NATIVE_ENDIAN, + .impl = { + .min_access_size = 4, + .max_access_size = 4, + .unaligned = false, + }, + .valid = { + .min_access_size = 4, + .max_access_size = 4, + .unaligned = false, + }, +}; + +static void stm32l4x5_gpio_init(Object *obj) +{ + Stm32l4x5GpioState *s = STM32L4X5_GPIO(obj); + + memory_region_init_io(&s->mmio, obj, &stm32l4x5_gpio_ops, s, + TYPE_STM32L4X5_GPIO, 0x400); + + sysbus_init_mmio(SYS_BUS_DEVICE(obj), &s->mmio); + + qdev_init_gpio_out(DEVICE(obj), s->pin, GPIO_NUM_PINS); + qdev_init_gpio_in(DEVICE(obj), stm32l4x5_gpio_set, GPIO_NUM_PINS); + + s->clk = qdev_init_clock_in(DEVICE(s), "clk", NULL, s, 0); + + object_property_add(obj, "disconnected-pins", "uint16", + disconnected_pins_get, disconnected_pins_set, + NULL, &s->disconnected_pins); + object_property_add(obj, "clock-freq-hz", "uint32", + clock_freq_get, NULL, NULL, NULL); +} + +static void stm32l4x5_gpio_realize(DeviceState *dev, Error **errp) +{ + Stm32l4x5GpioState *s = STM32L4X5_GPIO(dev); + if (!clock_has_source(s->clk)) { + error_setg(errp, "GPIO: clk input must be connected"); + return; + } +} + +static const VMStateDescription vmstate_stm32l4x5_gpio = { + .name = TYPE_STM32L4X5_GPIO, + .version_id = 1, + .minimum_version_id = 1, + .fields = (VMStateField[]){ + VMSTATE_UINT32(moder, Stm32l4x5GpioState), + VMSTATE_UINT32(otyper, Stm32l4x5GpioState), + VMSTATE_UINT32(ospeedr, Stm32l4x5GpioState), + VMSTATE_UINT32(pupdr, Stm32l4x5GpioState), + VMSTATE_UINT32(idr, Stm32l4x5GpioState), + VMSTATE_UINT32(odr, Stm32l4x5GpioState), + VMSTATE_UINT32(lckr, Stm32l4x5GpioState), + VMSTATE_UINT32(afrl, Stm32l4x5GpioState), + VMSTATE_UINT32(afrh, Stm32l4x5GpioState), + VMSTATE_UINT32(ascr, Stm32l4x5GpioState), + VMSTATE_UINT16(disconnected_pins, Stm32l4x5GpioState), + VMSTATE_UINT16(pins_connected_high, Stm32l4x5GpioState), + VMSTATE_END_OF_LIST() + } +}; + +static Property stm32l4x5_gpio_properties[] = { + DEFINE_PROP_STRING("name", Stm32l4x5GpioState, name), + DEFINE_PROP_UINT32("mode-reset", Stm32l4x5GpioState, moder_reset, 0), + DEFINE_PROP_UINT32("ospeed-reset", Stm32l4x5GpioState, ospeedr_reset, 0), + DEFINE_PROP_UINT32("pupd-reset", Stm32l4x5GpioState, pupdr_reset, 0), + DEFINE_PROP_END_OF_LIST(), +}; + +static void stm32l4x5_gpio_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + ResettableClass *rc = RESETTABLE_CLASS(klass); + + device_class_set_props(dc, stm32l4x5_gpio_properties); + dc->vmsd = &vmstate_stm32l4x5_gpio; + dc->realize = stm32l4x5_gpio_realize; + rc->phases.hold = stm32l4x5_gpio_reset_hold; +} + +static const TypeInfo stm32l4x5_gpio_types[] = { + { + .name = TYPE_STM32L4X5_GPIO, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(Stm32l4x5GpioState), + .instance_init = stm32l4x5_gpio_init, + .class_init = stm32l4x5_gpio_class_init, + }, +}; + +DEFINE_TYPES(stm32l4x5_gpio_types) diff --git a/hw/gpio/Kconfig b/hw/gpio/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/gpio/Kconfig +++ b/hw/gpio/Kconfig @@ -XXX,XX +XXX,XX @@ config GPIO_PWR config SIFIVE_GPIO bool + +config STM32L4X5_GPIO + bool diff --git a/hw/gpio/meson.build b/hw/gpio/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/gpio/meson.build +++ b/hw/gpio/meson.build @@ -XXX,XX +XXX,XX @@ system_ss.add(when: 'CONFIG_RASPI', if_true: files( 'bcm2835_gpio.c', 'bcm2838_gpio.c' )) +system_ss.add(when: 'CONFIG_STM32L4X5_SOC', if_true: files('stm32l4x5_gpio.c')) system_ss.add(when: 'CONFIG_ASPEED_SOC', if_true: files('aspeed_gpio.c')) system_ss.add(when: 'CONFIG_SIFIVE_GPIO', if_true: files('sifive_gpio.c')) diff --git a/hw/gpio/trace-events b/hw/gpio/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/gpio/trace-events +++ b/hw/gpio/trace-events @@ -XXX,XX +XXX,XX @@ sifive_gpio_update_output_irq(int64_t line, int64_t value) "line %" PRIi64 " val # aspeed_gpio.c aspeed_gpio_read(uint64_t offset, uint64_t value) "offset: 0x%" PRIx64 " value 0x%" PRIx64 aspeed_gpio_write(uint64_t offset, uint64_t value) "offset: 0x%" PRIx64 " value 0x%" PRIx64 + +# stm32l4x5_gpio.c +stm32l4x5_gpio_read(char *gpio, uint64_t addr) "GPIO%s addr: 0x%" PRIx64 " " +stm32l4x5_gpio_write(char *gpio, uint64_t addr, uint64_t data) "GPIO%s addr: 0x%" PRIx64 " val: 0x%" PRIx64 "" +stm32l4x5_gpio_update_idr(char *gpio, uint32_t old_idr, uint32_t new_idr) "GPIO%s from: 0x%x to: 0x%x" +stm32l4x5_gpio_pins(char *gpio, uint16_t disconnected, uint16_t high) "GPIO%s disconnected pins: 0x%x levels: 0x%x" -- 2.34.1 From: Inès Varhol <ines.varhol@telecom-paris.fr> Signed-off-by: Arnaud Minier <arnaud.minier@telecom-paris.fr> Signed-off-by: Inès Varhol <ines.varhol@telecom-paris.fr> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Acked-by: Alistair Francis <alistair.francis@wdc.com> Message-id: 20240305210444.310665-3-ines.varhol@telecom-paris.fr Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/stm32l4x5_soc.h | 2 + include/hw/gpio/stm32l4x5_gpio.h | 1 + include/hw/misc/stm32l4x5_syscfg.h | 3 +- hw/arm/stm32l4x5_soc.c | 71 +++++++++++++++++++++++------- hw/misc/stm32l4x5_syscfg.c | 1 + hw/arm/Kconfig | 3 +- 6 files changed, 63 insertions(+), 18 deletions(-) diff --git a/include/hw/arm/stm32l4x5_soc.h b/include/hw/arm/stm32l4x5_soc.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/stm32l4x5_soc.h +++ b/include/hw/arm/stm32l4x5_soc.h @@ -XXX,XX +XXX,XX @@ #include "hw/misc/stm32l4x5_syscfg.h" #include "hw/misc/stm32l4x5_exti.h" #include "hw/misc/stm32l4x5_rcc.h" +#include "hw/gpio/stm32l4x5_gpio.h" #include "qom/object.h" #define TYPE_STM32L4X5_SOC "stm32l4x5-soc" @@ -XXX,XX +XXX,XX @@ struct Stm32l4x5SocState { OrIRQState exti_or_gates[NUM_EXTI_OR_GATES]; Stm32l4x5SyscfgState syscfg; Stm32l4x5RccState rcc; + Stm32l4x5GpioState gpio[NUM_GPIOS]; MemoryRegion sram1; MemoryRegion sram2; diff --git a/include/hw/gpio/stm32l4x5_gpio.h b/include/hw/gpio/stm32l4x5_gpio.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/gpio/stm32l4x5_gpio.h +++ b/include/hw/gpio/stm32l4x5_gpio.h @@ -XXX,XX +XXX,XX @@ #define TYPE_STM32L4X5_GPIO "stm32l4x5-gpio" OBJECT_DECLARE_SIMPLE_TYPE(Stm32l4x5GpioState, STM32L4X5_GPIO) +#define NUM_GPIOS 8 #define GPIO_NUM_PINS 16 struct Stm32l4x5GpioState { diff --git a/include/hw/misc/stm32l4x5_syscfg.h b/include/hw/misc/stm32l4x5_syscfg.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/stm32l4x5_syscfg.h +++ b/include/hw/misc/stm32l4x5_syscfg.h @@ -XXX,XX +XXX,XX @@ #include "hw/sysbus.h" #include "qom/object.h" +#include "hw/gpio/stm32l4x5_gpio.h" #define TYPE_STM32L4X5_SYSCFG "stm32l4x5-syscfg" OBJECT_DECLARE_SIMPLE_TYPE(Stm32l4x5SyscfgState, STM32L4X5_SYSCFG) -#define NUM_GPIOS 8 -#define GPIO_NUM_PINS 16 #define SYSCFG_NUM_EXTICR 4 struct Stm32l4x5SyscfgState { diff --git a/hw/arm/stm32l4x5_soc.c b/hw/arm/stm32l4x5_soc.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/stm32l4x5_soc.c +++ b/hw/arm/stm32l4x5_soc.c @@ -XXX,XX +XXX,XX @@ #include "sysemu/sysemu.h" #include "hw/or-irq.h" #include "hw/arm/stm32l4x5_soc.h" +#include "hw/gpio/stm32l4x5_gpio.h" #include "hw/qdev-clock.h" #include "hw/misc/unimp.h" @@ -XXX,XX +XXX,XX @@ static const int exti_or_gate1_lines_in[EXTI_OR_GATE1_NUM_LINES_IN] = { 16, 35, 36, 37, 38, }; +static const struct { + uint32_t addr; + uint32_t moder_reset; + uint32_t ospeedr_reset; + uint32_t pupdr_reset; +} stm32l4x5_gpio_cfg[NUM_GPIOS] = { + { 0x48000000, 0xABFFFFFF, 0x0C000000, 0x64000000 }, + { 0x48000400, 0xFFFFFEBF, 0x00000000, 0x00000100 }, + { 0x48000800, 0xFFFFFFFF, 0x00000000, 0x00000000 }, + { 0x48000C00, 0xFFFFFFFF, 0x00000000, 0x00000000 }, + { 0x48001000, 0xFFFFFFFF, 0x00000000, 0x00000000 }, + { 0x48001400, 0xFFFFFFFF, 0x00000000, 0x00000000 }, + { 0x48001800, 0xFFFFFFFF, 0x00000000, 0x00000000 }, + { 0x48001C00, 0x0000000F, 0x00000000, 0x00000000 }, +}; + static void stm32l4x5_soc_initfn(Object *obj) { Stm32l4x5SocState *s = STM32L4X5_SOC(obj); @@ -XXX,XX +XXX,XX @@ static void stm32l4x5_soc_initfn(Object *obj) } object_initialize_child(obj, "syscfg", &s->syscfg, TYPE_STM32L4X5_SYSCFG); object_initialize_child(obj, "rcc", &s->rcc, TYPE_STM32L4X5_RCC); + + for (unsigned i = 0; i < NUM_GPIOS; i++) { + g_autofree char *name = g_strdup_printf("gpio%c", 'a' + i); + object_initialize_child(obj, name, &s->gpio[i], TYPE_STM32L4X5_GPIO); + } } static void stm32l4x5_soc_realize(DeviceState *dev_soc, Error **errp) @@ -XXX,XX +XXX,XX @@ static void stm32l4x5_soc_realize(DeviceState *dev_soc, Error **errp) Stm32l4x5SocState *s = STM32L4X5_SOC(dev_soc); const Stm32l4x5SocClass *sc = STM32L4X5_SOC_GET_CLASS(dev_soc); MemoryRegion *system_memory = get_system_memory(); - DeviceState *armv7m; + DeviceState *armv7m, *dev; SysBusDevice *busdev; + uint32_t pin_index; if (!memory_region_init_rom(&s->flash, OBJECT(dev_soc), "flash", sc->flash_size, errp)) { @@ -XXX,XX +XXX,XX @@ static void stm32l4x5_soc_realize(DeviceState *dev_soc, Error **errp) return; } + /* GPIOs */ + for (unsigned i = 0; i < NUM_GPIOS; i++) { + g_autofree char *name = g_strdup_printf("%c", 'A' + i); + dev = DEVICE(&s->gpio[i]); + qdev_prop_set_string(dev, "name", name); + qdev_prop_set_uint32(dev, "mode-reset", + stm32l4x5_gpio_cfg[i].moder_reset); + qdev_prop_set_uint32(dev, "ospeed-reset", + stm32l4x5_gpio_cfg[i].ospeedr_reset); + qdev_prop_set_uint32(dev, "pupd-reset", + stm32l4x5_gpio_cfg[i].pupdr_reset); + busdev = SYS_BUS_DEVICE(&s->gpio[i]); + g_free(name); + name = g_strdup_printf("gpio%c-out", 'a' + i); + qdev_connect_clock_in(DEVICE(&s->gpio[i]), "clk", + qdev_get_clock_out(DEVICE(&(s->rcc)), name)); + if (!sysbus_realize(busdev, errp)) { + return; + } + sysbus_mmio_map(busdev, 0, stm32l4x5_gpio_cfg[i].addr); + } + /* System configuration controller */ busdev = SYS_BUS_DEVICE(&s->syscfg); if (!sysbus_realize(busdev, errp)) { return; } sysbus_mmio_map(busdev, 0, SYSCFG_ADDR); - /* - * TODO: when the GPIO device is implemented, connect it - * to SYCFG using `qdev_connect_gpio_out`, NUM_GPIOS and - * GPIO_NUM_PINS. - */ + + for (unsigned i = 0; i < NUM_GPIOS; i++) { + for (unsigned j = 0; j < GPIO_NUM_PINS; j++) { + pin_index = GPIO_NUM_PINS * i + j; + qdev_connect_gpio_out(DEVICE(&s->gpio[i]), j, + qdev_get_gpio_in(DEVICE(&s->syscfg), + pin_index)); + } + } /* EXTI device */ busdev = SYS_BUS_DEVICE(&s->exti); @@ -XXX,XX +XXX,XX @@ static void stm32l4x5_soc_realize(DeviceState *dev_soc, Error **errp) } } - for (unsigned i = 0; i < 16; i++) { + for (unsigned i = 0; i < GPIO_NUM_PINS; i++) { qdev_connect_gpio_out(DEVICE(&s->syscfg), i, qdev_get_gpio_in(DEVICE(&s->exti), i)); } @@ -XXX,XX +XXX,XX @@ static void stm32l4x5_soc_realize(DeviceState *dev_soc, Error **errp) /* RESERVED: 0x40024400, 0x7FDBC00 */ /* AHB2 BUS */ - create_unimplemented_device("GPIOA", 0x48000000, 0x400); - create_unimplemented_device("GPIOB", 0x48000400, 0x400); - create_unimplemented_device("GPIOC", 0x48000800, 0x400); - create_unimplemented_device("GPIOD", 0x48000C00, 0x400); - create_unimplemented_device("GPIOE", 0x48001000, 0x400); - create_unimplemented_device("GPIOF", 0x48001400, 0x400); - create_unimplemented_device("GPIOG", 0x48001800, 0x400); - create_unimplemented_device("GPIOH", 0x48001C00, 0x400); /* RESERVED: 0x48002000, 0x7FDBC00 */ create_unimplemented_device("OTG_FS", 0x50000000, 0x40000); create_unimplemented_device("ADC", 0x50040000, 0x400); diff --git a/hw/misc/stm32l4x5_syscfg.c b/hw/misc/stm32l4x5_syscfg.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/stm32l4x5_syscfg.c +++ b/hw/misc/stm32l4x5_syscfg.c @@ -XXX,XX +XXX,XX @@ #include "hw/irq.h" #include "migration/vmstate.h" #include "hw/misc/stm32l4x5_syscfg.h" +#include "hw/gpio/stm32l4x5_gpio.h" #define SYSCFG_MEMRMP 0x00 #define SYSCFG_CFGR1 0x04 diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Kconfig +++ b/hw/arm/Kconfig @@ -XXX,XX +XXX,XX @@ config STM32L4X5_SOC bool select ARM_V7M select OR_IRQ - select STM32L4X5_SYSCFG select STM32L4X5_EXTI + select STM32L4X5_SYSCFG select STM32L4X5_RCC + select STM32L4X5_GPIO config XLNX_ZYNQMP_ARM bool -- 2.34.1 From: Inès Varhol <ines.varhol@telecom-paris.fr> The testcase contains : - `test_idr_reset_value()` : Checks the reset values of MODER, OTYPER, PUPDR, ODR and IDR. - `test_gpio_output_mode()` : Checks that writing a bit in register ODR results in the corresponding pin rising or lowering, if this pin is configured in output mode. - `test_gpio_input_mode()` : Checks that a input pin set high or low externally results in the pin rising and lowering. - `test_pull_up_pull_down()` : Checks that a floating pin in pull-up/down mode is actually high/down. - `test_push_pull()` : Checks that a pin set externally is disconnected when configured in push-pull output mode, and can't be set externally while in this mode. - `test_open_drain()` : Checks that a pin set externally high is disconnected when configured in open-drain output mode, and can't be set high while in this mode. - `test_bsrr_brr()` : Checks that writing to BSRR and BRR has the desired result in ODR. - `test_clock_enable()` : Checks that GPIO clock is at the right frequency after enabling it. Acked-by: Thomas Huth <thuth@redhat.com> Signed-off-by: Arnaud Minier <arnaud.minier@telecom-paris.fr> Signed-off-by: Inès Varhol <ines.varhol@telecom-paris.fr> Message-id: 20240305210444.310665-4-ines.varhol@telecom-paris.fr Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- tests/qtest/stm32l4x5_gpio-test.c | 551 ++++++++++++++++++++++++++++++ tests/qtest/meson.build | 3 +- 2 files changed, 553 insertions(+), 1 deletion(-) create mode 100644 tests/qtest/stm32l4x5_gpio-test.c diff --git a/tests/qtest/stm32l4x5_gpio-test.c b/tests/qtest/stm32l4x5_gpio-test.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/qtest/stm32l4x5_gpio-test.c @@ -XXX,XX +XXX,XX @@ +/* + * QTest testcase for STM32L4x5_GPIO + * + * Copyright (c) 2024 Arnaud Minier <arnaud.minier@telecom-paris.fr> + * Copyright (c) 2024 Inès Varhol <ines.varhol@telecom-paris.fr> + * + * This work is licensed under the terms of the GNU GPL, version 2 or later. + * See the COPYING file in the top-level directory. + */ + +#include "qemu/osdep.h" +#include "libqtest-single.h" + +#define GPIO_BASE_ADDR 0x48000000 +#define GPIO_SIZE 0x400 +#define NUM_GPIOS 8 +#define NUM_GPIO_PINS 16 + +#define GPIO_A 0x48000000 +#define GPIO_B 0x48000400 +#define GPIO_C 0x48000800 +#define GPIO_D 0x48000C00 +#define GPIO_E 0x48001000 +#define GPIO_F 0x48001400 +#define GPIO_G 0x48001800 +#define GPIO_H 0x48001C00 + +#define MODER 0x00 +#define OTYPER 0x04 +#define PUPDR 0x0C +#define IDR 0x10 +#define ODR 0x14 +#define BSRR 0x18 +#define BRR 0x28 + +#define MODER_INPUT 0 +#define MODER_OUTPUT 1 + +#define PUPDR_NONE 0 +#define PUPDR_PULLUP 1 +#define PUPDR_PULLDOWN 2 + +#define OTYPER_PUSH_PULL 0 +#define OTYPER_OPEN_DRAIN 1 + +const uint32_t moder_reset[NUM_GPIOS] = { + 0xABFFFFFF, + 0xFFFFFEBF, + 0xFFFFFFFF, + 0xFFFFFFFF, + 0xFFFFFFFF, + 0xFFFFFFFF, + 0xFFFFFFFF, + 0x0000000F +}; + +const uint32_t pupdr_reset[NUM_GPIOS] = { + 0x64000000, + 0x00000100, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000 +}; + +const uint32_t idr_reset[NUM_GPIOS] = { + 0x0000A000, + 0x00000010, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000 +}; + +static uint32_t gpio_readl(unsigned int gpio, unsigned int offset) +{ + return readl(gpio + offset); +} + +static void gpio_writel(unsigned int gpio, unsigned int offset, uint32_t value) +{ + writel(gpio + offset, value); +} + +static void gpio_set_bit(unsigned int gpio, unsigned int reg, + unsigned int pin, uint32_t value) +{ + uint32_t mask = 0xFFFFFFFF & ~(0x1 << pin); + gpio_writel(gpio, reg, (gpio_readl(gpio, reg) & mask) | value << pin); +} + +static void gpio_set_2bits(unsigned int gpio, unsigned int reg, + unsigned int pin, uint32_t value) +{ + uint32_t offset = 2 * pin; + uint32_t mask = 0xFFFFFFFF & ~(0x3 << offset); + gpio_writel(gpio, reg, (gpio_readl(gpio, reg) & mask) | value << offset); +} + +static unsigned int get_gpio_id(uint32_t gpio_addr) +{ + return (gpio_addr - GPIO_BASE_ADDR) / GPIO_SIZE; +} + +static void gpio_set_irq(unsigned int gpio, int num, int level) +{ + g_autofree char *name = g_strdup_printf("/machine/soc/gpio%c", + get_gpio_id(gpio) + 'a'); + qtest_set_irq_in(global_qtest, name, NULL, num, level); +} + +static void disconnect_all_pins(unsigned int gpio) +{ + g_autofree char *path = g_strdup_printf("/machine/soc/gpio%c", + get_gpio_id(gpio) + 'a'); + QDict *r; + + r = qtest_qmp(global_qtest, "{ 'execute': 'qom-set', 'arguments': " + "{ 'path': %s, 'property': 'disconnected-pins', 'value': %d } }", + path, 0xFFFF); + g_assert_false(qdict_haskey(r, "error")); + qobject_unref(r); +} + +static uint32_t get_disconnected_pins(unsigned int gpio) +{ + g_autofree char *path = g_strdup_printf("/machine/soc/gpio%c", + get_gpio_id(gpio) + 'a'); + uint32_t disconnected_pins = 0; + QDict *r; + + r = qtest_qmp(global_qtest, "{ 'execute': 'qom-get', 'arguments':" + " { 'path': %s, 'property': 'disconnected-pins'} }", path); + g_assert_false(qdict_haskey(r, "error")); + disconnected_pins = qdict_get_int(r, "return"); + qobject_unref(r); + return disconnected_pins; +} + +static uint32_t reset(uint32_t gpio, unsigned int offset) +{ + switch (offset) { + case MODER: + return moder_reset[get_gpio_id(gpio)]; + case PUPDR: + return pupdr_reset[get_gpio_id(gpio)]; + case IDR: + return idr_reset[get_gpio_id(gpio)]; + } + return 0x0; +} + +static void system_reset(void) +{ + QDict *r; + r = qtest_qmp(global_qtest, "{'execute': 'system_reset'}"); + g_assert_false(qdict_haskey(r, "error")); + qobject_unref(r); +} + +static void test_idr_reset_value(void) +{ + /* + * Checks that the values in MODER, OTYPER, PUPDR and ODR + * after reset are correct, and that the value in IDR is + * coherent. + * Since AF and analog modes aren't implemented, IDR reset + * values aren't the same as with a real board. + * + * Register IDR contains the actual values of all GPIO pins. + * Its value depends on the pins' configuration + * (intput/output/analog : register MODER, push-pull/open-drain : + * register OTYPER, pull-up/pull-down/none : register PUPDR) + * and on the values stored in register ODR + * (in case the pin is in output mode). + */ + + gpio_writel(GPIO_A, MODER, 0xDEADBEEF); + gpio_writel(GPIO_A, ODR, 0xDEADBEEF); + gpio_writel(GPIO_A, OTYPER, 0xDEADBEEF); + gpio_writel(GPIO_A, PUPDR, 0xDEADBEEF); + + gpio_writel(GPIO_B, MODER, 0xDEADBEEF); + gpio_writel(GPIO_B, ODR, 0xDEADBEEF); + gpio_writel(GPIO_B, OTYPER, 0xDEADBEEF); + gpio_writel(GPIO_B, PUPDR, 0xDEADBEEF); + + gpio_writel(GPIO_C, MODER, 0xDEADBEEF); + gpio_writel(GPIO_C, ODR, 0xDEADBEEF); + gpio_writel(GPIO_C, OTYPER, 0xDEADBEEF); + gpio_writel(GPIO_C, PUPDR, 0xDEADBEEF); + + gpio_writel(GPIO_H, MODER, 0xDEADBEEF); + gpio_writel(GPIO_H, ODR, 0xDEADBEEF); + gpio_writel(GPIO_H, OTYPER, 0xDEADBEEF); + gpio_writel(GPIO_H, PUPDR, 0xDEADBEEF); + + system_reset(); + + uint32_t moder = gpio_readl(GPIO_A, MODER); + uint32_t odr = gpio_readl(GPIO_A, ODR); + uint32_t otyper = gpio_readl(GPIO_A, OTYPER); + uint32_t pupdr = gpio_readl(GPIO_A, PUPDR); + uint32_t idr = gpio_readl(GPIO_A, IDR); + /* 15: AF, 14: AF, 13: AF, 12: Analog ... */ + /* here AF is the same as Analog and Input mode */ + g_assert_cmphex(moder, ==, reset(GPIO_A, MODER)); + g_assert_cmphex(odr, ==, reset(GPIO_A, ODR)); + g_assert_cmphex(otyper, ==, reset(GPIO_A, OTYPER)); + /* 15: pull-up, 14: pull-down, 13: pull-up, 12: neither ... */ + g_assert_cmphex(pupdr, ==, reset(GPIO_A, PUPDR)); + /* 15 : 1, 14: 0, 13: 1, 12 : reset value ... */ + g_assert_cmphex(idr, ==, reset(GPIO_A, IDR)); + + moder = gpio_readl(GPIO_B, MODER); + odr = gpio_readl(GPIO_B, ODR); + otyper = gpio_readl(GPIO_B, OTYPER); + pupdr = gpio_readl(GPIO_B, PUPDR); + idr = gpio_readl(GPIO_B, IDR); + /* ... 5: Analog, 4: AF, 3: AF, 2: Analog ... */ + /* here AF is the same as Analog and Input mode */ + g_assert_cmphex(moder, ==, reset(GPIO_B, MODER)); + g_assert_cmphex(odr, ==, reset(GPIO_B, ODR)); + g_assert_cmphex(otyper, ==, reset(GPIO_B, OTYPER)); + /* ... 5: neither, 4: pull-up, 3: neither ... */ + g_assert_cmphex(pupdr, ==, reset(GPIO_B, PUPDR)); + /* ... 5 : reset value, 4 : 1, 3 : reset value ... */ + g_assert_cmphex(idr, ==, reset(GPIO_B, IDR)); + + moder = gpio_readl(GPIO_C, MODER); + odr = gpio_readl(GPIO_C, ODR); + otyper = gpio_readl(GPIO_C, OTYPER); + pupdr = gpio_readl(GPIO_C, PUPDR); + idr = gpio_readl(GPIO_C, IDR); + /* Analog, same as Input mode*/ + g_assert_cmphex(moder, ==, reset(GPIO_C, MODER)); + g_assert_cmphex(odr, ==, reset(GPIO_C, ODR)); + g_assert_cmphex(otyper, ==, reset(GPIO_C, OTYPER)); + /* no pull-up or pull-down */ + g_assert_cmphex(pupdr, ==, reset(GPIO_C, PUPDR)); + /* reset value */ + g_assert_cmphex(idr, ==, reset(GPIO_C, IDR)); + + moder = gpio_readl(GPIO_H, MODER); + odr = gpio_readl(GPIO_H, ODR); + otyper = gpio_readl(GPIO_H, OTYPER); + pupdr = gpio_readl(GPIO_H, PUPDR); + idr = gpio_readl(GPIO_H, IDR); + /* Analog, same as Input mode */ + g_assert_cmphex(moder, ==, reset(GPIO_H, MODER)); + g_assert_cmphex(odr, ==, reset(GPIO_H, ODR)); + g_assert_cmphex(otyper, ==, reset(GPIO_H, OTYPER)); + /* no pull-up or pull-down */ + g_assert_cmphex(pupdr, ==, reset(GPIO_H, PUPDR)); + /* reset value */ + g_assert_cmphex(idr, ==, reset(GPIO_H, IDR)); +} + +static void test_gpio_output_mode(const void *data) +{ + /* + * Checks that setting a bit in ODR sets the corresponding + * GPIO line high : it should set the right bit in IDR + * and send an irq to syscfg. + * Additionally, it checks that values written to ODR + * when not in output mode are stored and not discarded. + */ + unsigned int pin = ((uint64_t)data) & 0xF; + uint32_t gpio = ((uint64_t)data) >> 32; + unsigned int gpio_id = get_gpio_id(gpio); + + qtest_irq_intercept_in(global_qtest, "/machine/soc/syscfg"); + + /* Set a bit in ODR and check nothing happens */ + gpio_set_bit(gpio, ODR, pin, 1); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR)); + g_assert_false(get_irq(gpio_id * NUM_GPIO_PINS + pin)); + + /* Configure the relevant line as output and check the pin is high */ + gpio_set_2bits(gpio, MODER, pin, MODER_OUTPUT); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR) | (1 << pin)); + g_assert_true(get_irq(gpio_id * NUM_GPIO_PINS + pin)); + + /* Reset the bit in ODR and check the pin is low */ + gpio_set_bit(gpio, ODR, pin, 0); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR) & ~(1 << pin)); + g_assert_false(get_irq(gpio_id * NUM_GPIO_PINS + pin)); + + /* Clean the test */ + gpio_writel(gpio, ODR, reset(gpio, ODR)); + gpio_writel(gpio, MODER, reset(gpio, MODER)); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR)); + g_assert_false(get_irq(gpio_id * NUM_GPIO_PINS + pin)); +} + +static void test_gpio_input_mode(const void *data) +{ + /* + * Test that setting a line high/low externally sets the + * corresponding GPIO line high/low : it should set the + * right bit in IDR and send an irq to syscfg. + */ + unsigned int pin = ((uint64_t)data) & 0xF; + uint32_t gpio = ((uint64_t)data) >> 32; + unsigned int gpio_id = get_gpio_id(gpio); + + qtest_irq_intercept_in(global_qtest, "/machine/soc/syscfg"); + + /* Configure a line as input, raise it, and check that the pin is high */ + gpio_set_2bits(gpio, MODER, pin, MODER_INPUT); + gpio_set_irq(gpio, pin, 1); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR) | (1 << pin)); + g_assert_true(get_irq(gpio_id * NUM_GPIO_PINS + pin)); + + /* Lower the line and check that the pin is low */ + gpio_set_irq(gpio, pin, 0); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR) & ~(1 << pin)); + g_assert_false(get_irq(gpio_id * NUM_GPIO_PINS + pin)); + + /* Clean the test */ + gpio_writel(gpio, MODER, reset(gpio, MODER)); + disconnect_all_pins(gpio); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR)); +} + +static void test_pull_up_pull_down(const void *data) +{ + /* + * Test that a floating pin with pull-up sets the pin + * high and vice-versa. + */ + unsigned int pin = ((uint64_t)data) & 0xF; + uint32_t gpio = ((uint64_t)data) >> 32; + unsigned int gpio_id = get_gpio_id(gpio); + + qtest_irq_intercept_in(global_qtest, "/machine/soc/syscfg"); + + /* Configure a line as input with pull-up, check the line is set high */ + gpio_set_2bits(gpio, MODER, pin, MODER_INPUT); + gpio_set_2bits(gpio, PUPDR, pin, PUPDR_PULLUP); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR) | (1 << pin)); + g_assert_true(get_irq(gpio_id * NUM_GPIO_PINS + pin)); + + /* Configure the line with pull-down, check the line is low */ + gpio_set_2bits(gpio, PUPDR, pin, PUPDR_PULLDOWN); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR) & ~(1 << pin)); + g_assert_false(get_irq(gpio_id * NUM_GPIO_PINS + pin)); + + /* Clean the test */ + gpio_writel(gpio, MODER, reset(gpio, MODER)); + gpio_writel(gpio, PUPDR, reset(gpio, PUPDR)); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR)); +} + +static void test_push_pull(const void *data) +{ + /* + * Test that configuring a line in push-pull output mode + * disconnects the pin, that the pin can't be set or reset + * externally afterwards. + */ + unsigned int pin = ((uint64_t)data) & 0xF; + uint32_t gpio = ((uint64_t)data) >> 32; + uint32_t gpio2 = GPIO_BASE_ADDR + (GPIO_H - gpio); + + qtest_irq_intercept_in(global_qtest, "/machine/soc/syscfg"); + + /* Setting a line high externally, configuring it in push-pull output */ + /* And checking the pin was disconnected */ + gpio_set_irq(gpio, pin, 1); + gpio_set_2bits(gpio, MODER, pin, MODER_OUTPUT); + g_assert_cmphex(get_disconnected_pins(gpio), ==, 0xFFFF); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR) & ~(1 << pin)); + + /* Setting a line low externally, configuring it in push-pull output */ + /* And checking the pin was disconnected */ + gpio_set_irq(gpio2, pin, 0); + gpio_set_bit(gpio2, ODR, pin, 1); + gpio_set_2bits(gpio2, MODER, pin, MODER_OUTPUT); + g_assert_cmphex(get_disconnected_pins(gpio2), ==, 0xFFFF); + g_assert_cmphex(gpio_readl(gpio2, IDR), ==, reset(gpio2, IDR) | (1 << pin)); + + /* Trying to set a push-pull output pin, checking it doesn't work */ + gpio_set_irq(gpio, pin, 1); + g_assert_cmphex(get_disconnected_pins(gpio), ==, 0xFFFF); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR) & ~(1 << pin)); + + /* Trying to reset a push-pull output pin, checking it doesn't work */ + gpio_set_irq(gpio2, pin, 0); + g_assert_cmphex(get_disconnected_pins(gpio2), ==, 0xFFFF); + g_assert_cmphex(gpio_readl(gpio2, IDR), ==, reset(gpio2, IDR) | (1 << pin)); + + /* Clean the test */ + gpio_writel(gpio, MODER, reset(gpio, MODER)); + gpio_writel(gpio2, ODR, reset(gpio2, ODR)); + gpio_writel(gpio2, MODER, reset(gpio2, MODER)); +} + +static void test_open_drain(const void *data) +{ + /* + * Test that configuring a line in open-drain output mode + * disconnects a pin set high externally and that the pin + * can't be set high externally while configured in open-drain. + * + * However a pin set low externally shouldn't be disconnected, + * and it can be set low externally when in open-drain mode. + */ + unsigned int pin = ((uint64_t)data) & 0xF; + uint32_t gpio = ((uint64_t)data) >> 32; + uint32_t gpio2 = GPIO_BASE_ADDR + (GPIO_H - gpio); + + qtest_irq_intercept_in(global_qtest, "/machine/soc/syscfg"); + + /* Setting a line high externally, configuring it in open-drain output */ + /* And checking the pin was disconnected */ + gpio_set_irq(gpio, pin, 1); + gpio_set_bit(gpio, OTYPER, pin, OTYPER_OPEN_DRAIN); + gpio_set_2bits(gpio, MODER, pin, MODER_OUTPUT); + g_assert_cmphex(get_disconnected_pins(gpio), ==, 0xFFFF); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR) & ~(1 << pin)); + + /* Setting a line low externally, configuring it in open-drain output */ + /* And checking the pin wasn't disconnected */ + gpio_set_irq(gpio2, pin, 0); + gpio_set_bit(gpio2, ODR, pin, 1); + gpio_set_bit(gpio2, OTYPER, pin, OTYPER_OPEN_DRAIN); + gpio_set_2bits(gpio2, MODER, pin, MODER_OUTPUT); + g_assert_cmphex(get_disconnected_pins(gpio2), ==, 0xFFFF & ~(1 << pin)); + g_assert_cmphex(gpio_readl(gpio2, IDR), ==, + reset(gpio2, IDR) & ~(1 << pin)); + + /* Trying to set a open-drain output pin, checking it doesn't work */ + gpio_set_irq(gpio, pin, 1); + g_assert_cmphex(get_disconnected_pins(gpio), ==, 0xFFFF); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR) & ~(1 << pin)); + + /* Trying to reset a open-drain output pin, checking it works */ + gpio_set_bit(gpio, ODR, pin, 1); + gpio_set_irq(gpio, pin, 0); + g_assert_cmphex(get_disconnected_pins(gpio2), ==, 0xFFFF & ~(1 << pin)); + g_assert_cmphex(gpio_readl(gpio2, IDR), ==, + reset(gpio2, IDR) & ~(1 << pin)); + + /* Clean the test */ + disconnect_all_pins(gpio2); + gpio_writel(gpio2, OTYPER, reset(gpio2, OTYPER)); + gpio_writel(gpio2, ODR, reset(gpio2, ODR)); + gpio_writel(gpio2, MODER, reset(gpio2, MODER)); + g_assert_cmphex(gpio_readl(gpio2, IDR), ==, reset(gpio2, IDR)); + disconnect_all_pins(gpio); + gpio_writel(gpio, OTYPER, reset(gpio, OTYPER)); + gpio_writel(gpio, ODR, reset(gpio, ODR)); + gpio_writel(gpio, MODER, reset(gpio, MODER)); + g_assert_cmphex(gpio_readl(gpio, IDR), ==, reset(gpio, IDR)); +} + +static void test_bsrr_brr(const void *data) +{ + /* + * Test that writing a '1' in BSS and BSRR + * has the desired effect on ODR. + * In BSRR, BSx has priority over BRx. + */ + unsigned int pin = ((uint64_t)data) & 0xF; + uint32_t gpio = ((uint64_t)data) >> 32; + + gpio_writel(gpio, BSRR, (1 << pin)); + g_assert_cmphex(gpio_readl(gpio, ODR), ==, reset(gpio, ODR) | (1 << pin)); + + gpio_writel(gpio, BSRR, (1 << (pin + NUM_GPIO_PINS))); + g_assert_cmphex(gpio_readl(gpio, ODR), ==, reset(gpio, ODR)); + + gpio_writel(gpio, BSRR, (1 << pin)); + g_assert_cmphex(gpio_readl(gpio, ODR), ==, reset(gpio, ODR) | (1 << pin)); + + gpio_writel(gpio, BRR, (1 << pin)); + g_assert_cmphex(gpio_readl(gpio, ODR), ==, reset(gpio, ODR)); + + /* BSx should have priority over BRx */ + gpio_writel(gpio, BSRR, (1 << pin) | (1 << (pin + NUM_GPIO_PINS))); + g_assert_cmphex(gpio_readl(gpio, ODR), ==, reset(gpio, ODR) | (1 << pin)); + + gpio_writel(gpio, BRR, (1 << pin)); + g_assert_cmphex(gpio_readl(gpio, ODR), ==, reset(gpio, ODR)); + + gpio_writel(gpio, ODR, reset(gpio, ODR)); +} + +int main(int argc, char **argv) +{ + int ret; + + g_test_init(&argc, &argv, NULL); + g_test_set_nonfatal_assertions(); + qtest_add_func("stm32l4x5/gpio/test_idr_reset_value", + test_idr_reset_value); + /* + * The inputs for the tests (gpio and pin) can be changed, + * but the tests don't work for pins that are high at reset + * (GPIOA15, GPIO13 and GPIOB5). + * Specifically, rising the pin then checking `get_irq()` + * is problematic since the pin was already high. + */ + qtest_add_data_func("stm32l4x5/gpio/test_gpioc5_output_mode", + (void *)((uint64_t)GPIO_C << 32 | 5), + test_gpio_output_mode); + qtest_add_data_func("stm32l4x5/gpio/test_gpioh3_output_mode", + (void *)((uint64_t)GPIO_H << 32 | 3), + test_gpio_output_mode); + qtest_add_data_func("stm32l4x5/gpio/test_gpio_input_mode1", + (void *)((uint64_t)GPIO_D << 32 | 6), + test_gpio_input_mode); + qtest_add_data_func("stm32l4x5/gpio/test_gpio_input_mode2", + (void *)((uint64_t)GPIO_C << 32 | 10), + test_gpio_input_mode); + qtest_add_data_func("stm32l4x5/gpio/test_gpio_pull_up_pull_down1", + (void *)((uint64_t)GPIO_B << 32 | 5), + test_pull_up_pull_down); + qtest_add_data_func("stm32l4x5/gpio/test_gpio_pull_up_pull_down2", + (void *)((uint64_t)GPIO_F << 32 | 1), + test_pull_up_pull_down); + qtest_add_data_func("stm32l4x5/gpio/test_gpio_push_pull1", + (void *)((uint64_t)GPIO_G << 32 | 6), + test_push_pull); + qtest_add_data_func("stm32l4x5/gpio/test_gpio_push_pull2", + (void *)((uint64_t)GPIO_H << 32 | 3), + test_push_pull); + qtest_add_data_func("stm32l4x5/gpio/test_gpio_open_drain1", + (void *)((uint64_t)GPIO_C << 32 | 4), + test_open_drain); + qtest_add_data_func("stm32l4x5/gpio/test_gpio_open_drain2", + (void *)((uint64_t)GPIO_E << 32 | 11), + test_open_drain); + qtest_add_data_func("stm32l4x5/gpio/test_bsrr_brr1", + (void *)((uint64_t)GPIO_A << 32 | 12), + test_bsrr_brr); + qtest_add_data_func("stm32l4x5/gpio/test_bsrr_brr2", + (void *)((uint64_t)GPIO_D << 32 | 0), + test_bsrr_brr); + + qtest_start("-machine b-l475e-iot01a"); + ret = g_test_run(); + qtest_end(); + + return ret; +} diff --git a/tests/qtest/meson.build b/tests/qtest/meson.build index XXXXXXX..XXXXXXX 100644 --- a/tests/qtest/meson.build +++ b/tests/qtest/meson.build @@ -XXX,XX +XXX,XX @@ qtests_aspeed = \ qtests_stm32l4x5 = \ ['stm32l4x5_exti-test', 'stm32l4x5_syscfg-test', - 'stm32l4x5_rcc-test'] + 'stm32l4x5_rcc-test', + 'stm32l4x5_gpio-test'] qtests_arm = \ (config_all_devices.has_key('CONFIG_MPS2') ? ['sse-timer-test'] : []) + \ -- 2.34.1 From: Richard Henderson <richard.henderson@linaro.org> While the 8-bit input elements are sequential in the input vector, the 32-bit output elements are not sequential in the output matrix. Do not attempt to compute 2 32-bit outputs at the same time. Cc: qemu-stable@nongnu.org Fixes: 23a5e3859f5 ("target/arm: Implement SME integer outer product") Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2083 Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-id: 20240305163931.242795-1-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/tcg/sme_helper.c | 77 ++++++++++++++++++------------- tests/tcg/aarch64/sme-smopa-1.c | 47 +++++++++++++++++++ tests/tcg/aarch64/sme-smopa-2.c | 54 ++++++++++++++++++++++ tests/tcg/aarch64/Makefile.target | 2 +- 4 files changed, 147 insertions(+), 33 deletions(-) create mode 100644 tests/tcg/aarch64/sme-smopa-1.c create mode 100644 tests/tcg/aarch64/sme-smopa-2.c diff --git a/target/arm/tcg/sme_helper.c b/target/arm/tcg/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/sme_helper.c +++ b/target/arm/tcg/sme_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(sme_bfmopa)(void *vza, void *vzn, void *vzm, void *vpn, } } -typedef uint64_t IMOPFn(uint64_t, uint64_t, uint64_t, uint8_t, bool); +typedef uint32_t IMOPFn32(uint32_t, uint32_t, uint32_t, uint8_t, bool); +static inline void do_imopa_s(uint32_t *za, uint32_t *zn, uint32_t *zm, + uint8_t *pn, uint8_t *pm, + uint32_t desc, IMOPFn32 *fn) +{ + intptr_t row, col, oprsz = simd_oprsz(desc) / 4; + bool neg = simd_data(desc); -static inline void do_imopa(uint64_t *za, uint64_t *zn, uint64_t *zm, - uint8_t *pn, uint8_t *pm, - uint32_t desc, IMOPFn *fn) + for (row = 0; row < oprsz; ++row) { + uint8_t pa = (pn[H1(row >> 1)] >> ((row & 1) * 4)) & 0xf; + uint32_t *za_row = &za[tile_vslice_index(row)]; + uint32_t n = zn[H4(row)]; + + for (col = 0; col < oprsz; ++col) { + uint8_t pb = pm[H1(col >> 1)] >> ((col & 1) * 4); + uint32_t *a = &za_row[H4(col)]; + + *a = fn(n, zm[H4(col)], *a, pa & pb, neg); + } + } +} + +typedef uint64_t IMOPFn64(uint64_t, uint64_t, uint64_t, uint8_t, bool); +static inline void do_imopa_d(uint64_t *za, uint64_t *zn, uint64_t *zm, + uint8_t *pn, uint8_t *pm, + uint32_t desc, IMOPFn64 *fn) { intptr_t row, col, oprsz = simd_oprsz(desc) / 8; bool neg = simd_data(desc); @@ -XXX,XX +XXX,XX @@ static inline void do_imopa(uint64_t *za, uint64_t *zn, uint64_t *zm, } #define DEF_IMOP_32(NAME, NTYPE, MTYPE) \ -static uint64_t NAME(uint64_t n, uint64_t m, uint64_t a, uint8_t p, bool neg) \ +static uint32_t NAME(uint32_t n, uint32_t m, uint32_t a, uint8_t p, bool neg) \ { \ - uint32_t sum0 = 0, sum1 = 0; \ + uint32_t sum = 0; \ /* Apply P to N as a mask, making the inactive elements 0. */ \ n &= expand_pred_b(p); \ - sum0 += (NTYPE)(n >> 0) * (MTYPE)(m >> 0); \ - sum0 += (NTYPE)(n >> 8) * (MTYPE)(m >> 8); \ - sum0 += (NTYPE)(n >> 16) * (MTYPE)(m >> 16); \ - sum0 += (NTYPE)(n >> 24) * (MTYPE)(m >> 24); \ - sum1 += (NTYPE)(n >> 32) * (MTYPE)(m >> 32); \ - sum1 += (NTYPE)(n >> 40) * (MTYPE)(m >> 40); \ - sum1 += (NTYPE)(n >> 48) * (MTYPE)(m >> 48); \ - sum1 += (NTYPE)(n >> 56) * (MTYPE)(m >> 56); \ - if (neg) { \ - sum0 = (uint32_t)a - sum0, sum1 = (uint32_t)(a >> 32) - sum1; \ - } else { \ - sum0 = (uint32_t)a + sum0, sum1 = (uint32_t)(a >> 32) + sum1; \ - } \ - return ((uint64_t)sum1 << 32) | sum0; \ + sum += (NTYPE)(n >> 0) * (MTYPE)(m >> 0); \ + sum += (NTYPE)(n >> 8) * (MTYPE)(m >> 8); \ + sum += (NTYPE)(n >> 16) * (MTYPE)(m >> 16); \ + sum += (NTYPE)(n >> 24) * (MTYPE)(m >> 24); \ + return neg ? a - sum : a + sum; \ } #define DEF_IMOP_64(NAME, NTYPE, MTYPE) \ @@ -XXX,XX +XXX,XX @@ DEF_IMOP_64(umopa_d, uint16_t, uint16_t) DEF_IMOP_64(sumopa_d, int16_t, uint16_t) DEF_IMOP_64(usmopa_d, uint16_t, int16_t) -#define DEF_IMOPH(NAME) \ - void HELPER(sme_##NAME)(void *vza, void *vzn, void *vzm, void *vpn, \ - void *vpm, uint32_t desc) \ - { do_imopa(vza, vzn, vzm, vpn, vpm, desc, NAME); } +#define DEF_IMOPH(NAME, S) \ + void HELPER(sme_##NAME##_##S)(void *vza, void *vzn, void *vzm, \ + void *vpn, void *vpm, uint32_t desc) \ + { do_imopa_##S(vza, vzn, vzm, vpn, vpm, desc, NAME##_##S); } -DEF_IMOPH(smopa_s) -DEF_IMOPH(umopa_s) -DEF_IMOPH(sumopa_s) -DEF_IMOPH(usmopa_s) -DEF_IMOPH(smopa_d) -DEF_IMOPH(umopa_d) -DEF_IMOPH(sumopa_d) -DEF_IMOPH(usmopa_d) +DEF_IMOPH(smopa, s) +DEF_IMOPH(umopa, s) +DEF_IMOPH(sumopa, s) +DEF_IMOPH(usmopa, s) + +DEF_IMOPH(smopa, d) +DEF_IMOPH(umopa, d) +DEF_IMOPH(sumopa, d) +DEF_IMOPH(usmopa, d) diff --git a/tests/tcg/aarch64/sme-smopa-1.c b/tests/tcg/aarch64/sme-smopa-1.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/tcg/aarch64/sme-smopa-1.c @@ -XXX,XX +XXX,XX @@ +#include <stdio.h> +#include <string.h> + +int main() +{ + static const int cmp[4][4] = { + { 110, 134, 158, 182 }, + { 390, 478, 566, 654 }, + { 670, 822, 974, 1126 }, + { 950, 1166, 1382, 1598 } + }; + int dst[4][4]; + int *tmp = &dst[0][0]; + + asm volatile( + ".arch armv8-r+sme\n\t" + "smstart\n\t" + "index z0.b, #0, #1\n\t" + "movprfx z1, z0\n\t" + "add z1.b, z1.b, #16\n\t" + "ptrue p0.b\n\t" + "smopa za0.s, p0/m, p0/m, z0.b, z1.b\n\t" + "ptrue p0.s, vl4\n\t" + "mov w12, #0\n\t" + "st1w { za0h.s[w12, #0] }, p0, [%0]\n\t" + "add %0, %0, #16\n\t" + "st1w { za0h.s[w12, #1] }, p0, [%0]\n\t" + "add %0, %0, #16\n\t" + "st1w { za0h.s[w12, #2] }, p0, [%0]\n\t" + "add %0, %0, #16\n\t" + "st1w { za0h.s[w12, #3] }, p0, [%0]\n\t" + "smstop" + : "+r"(tmp) : : "memory"); + + if (memcmp(cmp, dst, sizeof(dst)) == 0) { + return 0; + } + + /* See above for correct results. */ + for (int i = 0; i < 4; ++i) { + for (int j = 0; j < 4; ++j) { + printf("%6d", dst[i][j]); + } + printf("\n"); + } + return 1; +} diff --git a/tests/tcg/aarch64/sme-smopa-2.c b/tests/tcg/aarch64/sme-smopa-2.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/tcg/aarch64/sme-smopa-2.c @@ -XXX,XX +XXX,XX @@ +#include <stdio.h> +#include <string.h> + +int main() +{ + static const long cmp[4][4] = { + { 110, 134, 158, 182 }, + { 390, 478, 566, 654 }, + { 670, 822, 974, 1126 }, + { 950, 1166, 1382, 1598 } + }; + long dst[4][4]; + long *tmp = &dst[0][0]; + long svl; + + /* Validate that we have a wide enough vector for 4 elements. */ + asm(".arch armv8-r+sme-i64\n\trdsvl %0, #1" : "=r"(svl)); + if (svl < 32) { + return 0; + } + + asm volatile( + "smstart\n\t" + "index z0.h, #0, #1\n\t" + "movprfx z1, z0\n\t" + "add z1.h, z1.h, #16\n\t" + "ptrue p0.b\n\t" + "smopa za0.d, p0/m, p0/m, z0.h, z1.h\n\t" + "ptrue p0.d, vl4\n\t" + "mov w12, #0\n\t" + "st1d { za0h.d[w12, #0] }, p0, [%0]\n\t" + "add %0, %0, #32\n\t" + "st1d { za0h.d[w12, #1] }, p0, [%0]\n\t" + "mov w12, #2\n\t" + "add %0, %0, #32\n\t" + "st1d { za0h.d[w12, #0] }, p0, [%0]\n\t" + "add %0, %0, #32\n\t" + "st1d { za0h.d[w12, #1] }, p0, [%0]\n\t" + "smstop" + : "+r"(tmp) : : "memory"); + + if (memcmp(cmp, dst, sizeof(dst)) == 0) { + return 0; + } + + /* See above for correct results. */ + for (int i = 0; i < 4; ++i) { + for (int j = 0; j < 4; ++j) { + printf("%6ld", dst[i][j]); + } + printf("\n"); + } + return 1; +} diff --git a/tests/tcg/aarch64/Makefile.target b/tests/tcg/aarch64/Makefile.target index XXXXXXX..XXXXXXX 100644 --- a/tests/tcg/aarch64/Makefile.target +++ b/tests/tcg/aarch64/Makefile.target @@ -XXX,XX +XXX,XX @@ endif # SME Tests ifneq ($(CROSS_AS_HAS_ARMV9_SME),) -AARCH64_TESTS += sme-outprod1 +AARCH64_TESTS += sme-outprod1 sme-smopa-1 sme-smopa-2 endif # System Registers Tests -- 2.34.1 The sun4v RTC device model added under commit a0e893039cf2ce0 in 2016 was unfortunately added with a license of GPL-v3-or-later, which is not compatible with other QEMU code which has a GPL-v2-only license. Relicense the code in the .c and the .h file to GPL-v2-or-later, to make it compatible with the rest of QEMU. Cc: qemu-stable@nongnu.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Paolo Bonzini (for Red Hat) <pbonzini@redhat.com> Signed-off-by: Artyom Tarasenko <atar4qemu@gmail.com> Signed-off-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> Acked-by: Alex Bennée <alex.bennee@linaro.org> Message-id: 20240223161300.938542-1-peter.maydell@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/rtc/sun4v-rtc.h | 2 +- hw/rtc/sun4v-rtc.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/include/hw/rtc/sun4v-rtc.h b/include/hw/rtc/sun4v-rtc.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/rtc/sun4v-rtc.h +++ b/include/hw/rtc/sun4v-rtc.h @@ -XXX,XX +XXX,XX @@ * * Copyright (c) 2016 Artyom Tarasenko * - * This code is licensed under the GNU GPL v3 or (at your option) any later + * This code is licensed under the GNU GPL v2 or (at your option) any later * version. */ diff --git a/hw/rtc/sun4v-rtc.c b/hw/rtc/sun4v-rtc.c index XXXXXXX..XXXXXXX 100644 --- a/hw/rtc/sun4v-rtc.c +++ b/hw/rtc/sun4v-rtc.c @@ -XXX,XX +XXX,XX @@ * * Copyright (c) 2016 Artyom Tarasenko * - * This code is licensed under the GNU GPL v3 or (at your option) any later + * This code is licensed under the GNU GPL v2 or (at your option) any later * version. */ -- 2.34.1 From: Thomas Huth <thuth@redhat.com> Move the code to a separate file so that we do not have to compile it anymore if CONFIG_ARM_V7M is not set. Signed-off-by: Thomas Huth <thuth@redhat.com> Message-id: 20240308141051.536599-2-thuth@redhat.com Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/tcg/cpu-v7m.c | 290 +++++++++++++++++++++++++++++++++++++ target/arm/tcg/cpu32.c | 261 --------------------------------- target/arm/meson.build | 3 + target/arm/tcg/meson.build | 3 + 4 files changed, 296 insertions(+), 261 deletions(-) create mode 100644 target/arm/tcg/cpu-v7m.c diff --git a/target/arm/tcg/cpu-v7m.c b/target/arm/tcg/cpu-v7m.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/tcg/cpu-v7m.c @@ -XXX,XX +XXX,XX @@ +/* + * QEMU ARMv7-M TCG-only CPUs. + * + * Copyright (c) 2012 SUSE LINUX Products GmbH + * + * This code is licensed under the GNU GPL v2 or later. + * + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#include "qemu/osdep.h" +#include "cpu.h" +#include "hw/core/tcg-cpu-ops.h" +#include "internals.h" + +#if !defined(CONFIG_USER_ONLY) + +#include "hw/intc/armv7m_nvic.h" + +static bool arm_v7m_cpu_exec_interrupt(CPUState *cs, int interrupt_request) +{ + CPUClass *cc = CPU_GET_CLASS(cs); + ARMCPU *cpu = ARM_CPU(cs); + CPUARMState *env = &cpu->env; + bool ret = false; + + /* + * ARMv7-M interrupt masking works differently than -A or -R. + * There is no FIQ/IRQ distinction. Instead of I and F bits + * masking FIQ and IRQ interrupts, an exception is taken only + * if it is higher priority than the current execution priority + * (which depends on state like BASEPRI, FAULTMASK and the + * currently active exception). + */ + if (interrupt_request & CPU_INTERRUPT_HARD + && (armv7m_nvic_can_take_pending_exception(env->nvic))) { + cs->exception_index = EXCP_IRQ; + cc->tcg_ops->do_interrupt(cs); + ret = true; + } + return ret; +} + +#endif /* !CONFIG_USER_ONLY */ + +static void cortex_m0_initfn(Object *obj) +{ + ARMCPU *cpu = ARM_CPU(obj); + set_feature(&cpu->env, ARM_FEATURE_V6); + set_feature(&cpu->env, ARM_FEATURE_M); + + cpu->midr = 0x410cc200; + + /* + * These ID register values are not guest visible, because + * we do not implement the Main Extension. They must be set + * to values corresponding to the Cortex-M0's implemented + * features, because QEMU generally controls its emulation + * by looking at ID register fields. We use the same values as + * for the M3. + */ + cpu->isar.id_pfr0 = 0x00000030; + cpu->isar.id_pfr1 = 0x00000200; + cpu->isar.id_dfr0 = 0x00100000; + cpu->id_afr0 = 0x00000000; + cpu->isar.id_mmfr0 = 0x00000030; + cpu->isar.id_mmfr1 = 0x00000000; + cpu->isar.id_mmfr2 = 0x00000000; + cpu->isar.id_mmfr3 = 0x00000000; + cpu->isar.id_isar0 = 0x01141110; + cpu->isar.id_isar1 = 0x02111000; + cpu->isar.id_isar2 = 0x21112231; + cpu->isar.id_isar3 = 0x01111110; + cpu->isar.id_isar4 = 0x01310102; + cpu->isar.id_isar5 = 0x00000000; + cpu->isar.id_isar6 = 0x00000000; +} + +static void cortex_m3_initfn(Object *obj) +{ + ARMCPU *cpu = ARM_CPU(obj); + set_feature(&cpu->env, ARM_FEATURE_V7); + set_feature(&cpu->env, ARM_FEATURE_M); + set_feature(&cpu->env, ARM_FEATURE_M_MAIN); + cpu->midr = 0x410fc231; + cpu->pmsav7_dregion = 8; + cpu->isar.id_pfr0 = 0x00000030; + cpu->isar.id_pfr1 = 0x00000200; + cpu->isar.id_dfr0 = 0x00100000; + cpu->id_afr0 = 0x00000000; + cpu->isar.id_mmfr0 = 0x00000030; + cpu->isar.id_mmfr1 = 0x00000000; + cpu->isar.id_mmfr2 = 0x00000000; + cpu->isar.id_mmfr3 = 0x00000000; + cpu->isar.id_isar0 = 0x01141110; + cpu->isar.id_isar1 = 0x02111000; + cpu->isar.id_isar2 = 0x21112231; + cpu->isar.id_isar3 = 0x01111110; + cpu->isar.id_isar4 = 0x01310102; + cpu->isar.id_isar5 = 0x00000000; + cpu->isar.id_isar6 = 0x00000000; +} + +static void cortex_m4_initfn(Object *obj) +{ + ARMCPU *cpu = ARM_CPU(obj); + + set_feature(&cpu->env, ARM_FEATURE_V7); + set_feature(&cpu->env, ARM_FEATURE_M); + set_feature(&cpu->env, ARM_FEATURE_M_MAIN); + set_feature(&cpu->env, ARM_FEATURE_THUMB_DSP); + cpu->midr = 0x410fc240; /* r0p0 */ + cpu->pmsav7_dregion = 8; + cpu->isar.mvfr0 = 0x10110021; + cpu->isar.mvfr1 = 0x11000011; + cpu->isar.mvfr2 = 0x00000000; + cpu->isar.id_pfr0 = 0x00000030; + cpu->isar.id_pfr1 = 0x00000200; + cpu->isar.id_dfr0 = 0x00100000; + cpu->id_afr0 = 0x00000000; + cpu->isar.id_mmfr0 = 0x00000030; + cpu->isar.id_mmfr1 = 0x00000000; + cpu->isar.id_mmfr2 = 0x00000000; + cpu->isar.id_mmfr3 = 0x00000000; + cpu->isar.id_isar0 = 0x01141110; + cpu->isar.id_isar1 = 0x02111000; + cpu->isar.id_isar2 = 0x21112231; + cpu->isar.id_isar3 = 0x01111110; + cpu->isar.id_isar4 = 0x01310102; + cpu->isar.id_isar5 = 0x00000000; + cpu->isar.id_isar6 = 0x00000000; +} + +static void cortex_m7_initfn(Object *obj) +{ + ARMCPU *cpu = ARM_CPU(obj); + + set_feature(&cpu->env, ARM_FEATURE_V7); + set_feature(&cpu->env, ARM_FEATURE_M); + set_feature(&cpu->env, ARM_FEATURE_M_MAIN); + set_feature(&cpu->env, ARM_FEATURE_THUMB_DSP); + cpu->midr = 0x411fc272; /* r1p2 */ + cpu->pmsav7_dregion = 8; + cpu->isar.mvfr0 = 0x10110221; + cpu->isar.mvfr1 = 0x12000011; + cpu->isar.mvfr2 = 0x00000040; + cpu->isar.id_pfr0 = 0x00000030; + cpu->isar.id_pfr1 = 0x00000200; + cpu->isar.id_dfr0 = 0x00100000; + cpu->id_afr0 = 0x00000000; + cpu->isar.id_mmfr0 = 0x00100030; + cpu->isar.id_mmfr1 = 0x00000000; + cpu->isar.id_mmfr2 = 0x01000000; + cpu->isar.id_mmfr3 = 0x00000000; + cpu->isar.id_isar0 = 0x01101110; + cpu->isar.id_isar1 = 0x02112000; + cpu->isar.id_isar2 = 0x20232231; + cpu->isar.id_isar3 = 0x01111131; + cpu->isar.id_isar4 = 0x01310132; + cpu->isar.id_isar5 = 0x00000000; + cpu->isar.id_isar6 = 0x00000000; +} + +static void cortex_m33_initfn(Object *obj) +{ + ARMCPU *cpu = ARM_CPU(obj); + + set_feature(&cpu->env, ARM_FEATURE_V8); + set_feature(&cpu->env, ARM_FEATURE_M); + set_feature(&cpu->env, ARM_FEATURE_M_MAIN); + set_feature(&cpu->env, ARM_FEATURE_M_SECURITY); + set_feature(&cpu->env, ARM_FEATURE_THUMB_DSP); + cpu->midr = 0x410fd213; /* r0p3 */ + cpu->pmsav7_dregion = 16; + cpu->sau_sregion = 8; + cpu->isar.mvfr0 = 0x10110021; + cpu->isar.mvfr1 = 0x11000011; + cpu->isar.mvfr2 = 0x00000040; + cpu->isar.id_pfr0 = 0x00000030; + cpu->isar.id_pfr1 = 0x00000210; + cpu->isar.id_dfr0 = 0x00200000; + cpu->id_afr0 = 0x00000000; + cpu->isar.id_mmfr0 = 0x00101F40; + cpu->isar.id_mmfr1 = 0x00000000; + cpu->isar.id_mmfr2 = 0x01000000; + cpu->isar.id_mmfr3 = 0x00000000; + cpu->isar.id_isar0 = 0x01101110; + cpu->isar.id_isar1 = 0x02212000; + cpu->isar.id_isar2 = 0x20232232; + cpu->isar.id_isar3 = 0x01111131; + cpu->isar.id_isar4 = 0x01310132; + cpu->isar.id_isar5 = 0x00000000; + cpu->isar.id_isar6 = 0x00000000; + cpu->clidr = 0x00000000; + cpu->ctr = 0x8000c000; +} + +static void cortex_m55_initfn(Object *obj) +{ + ARMCPU *cpu = ARM_CPU(obj); + + set_feature(&cpu->env, ARM_FEATURE_V8); + set_feature(&cpu->env, ARM_FEATURE_V8_1M); + set_feature(&cpu->env, ARM_FEATURE_M); + set_feature(&cpu->env, ARM_FEATURE_M_MAIN); + set_feature(&cpu->env, ARM_FEATURE_M_SECURITY); + set_feature(&cpu->env, ARM_FEATURE_THUMB_DSP); + cpu->midr = 0x410fd221; /* r0p1 */ + cpu->revidr = 0; + cpu->pmsav7_dregion = 16; + cpu->sau_sregion = 8; + /* These are the MVFR* values for the FPU + full MVE configuration */ + cpu->isar.mvfr0 = 0x10110221; + cpu->isar.mvfr1 = 0x12100211; + cpu->isar.mvfr2 = 0x00000040; + cpu->isar.id_pfr0 = 0x20000030; + cpu->isar.id_pfr1 = 0x00000230; + cpu->isar.id_dfr0 = 0x10200000; + cpu->id_afr0 = 0x00000000; + cpu->isar.id_mmfr0 = 0x00111040; + cpu->isar.id_mmfr1 = 0x00000000; + cpu->isar.id_mmfr2 = 0x01000000; + cpu->isar.id_mmfr3 = 0x00000011; + cpu->isar.id_isar0 = 0x01103110; + cpu->isar.id_isar1 = 0x02212000; + cpu->isar.id_isar2 = 0x20232232; + cpu->isar.id_isar3 = 0x01111131; + cpu->isar.id_isar4 = 0x01310132; + cpu->isar.id_isar5 = 0x00000000; + cpu->isar.id_isar6 = 0x00000000; + cpu->clidr = 0x00000000; /* caches not implemented */ + cpu->ctr = 0x8303c003; +} + +static const TCGCPUOps arm_v7m_tcg_ops = { + .initialize = arm_translate_init, + .synchronize_from_tb = arm_cpu_synchronize_from_tb, + .debug_excp_handler = arm_debug_excp_handler, + .restore_state_to_opc = arm_restore_state_to_opc, + +#ifdef CONFIG_USER_ONLY + .record_sigsegv = arm_cpu_record_sigsegv, + .record_sigbus = arm_cpu_record_sigbus, +#else + .tlb_fill = arm_cpu_tlb_fill, + .cpu_exec_interrupt = arm_v7m_cpu_exec_interrupt, + .do_interrupt = arm_v7m_cpu_do_interrupt, + .do_transaction_failed = arm_cpu_do_transaction_failed, + .do_unaligned_access = arm_cpu_do_unaligned_access, + .adjust_watchpoint_address = arm_adjust_watchpoint_address, + .debug_check_watchpoint = arm_debug_check_watchpoint, + .debug_check_breakpoint = arm_debug_check_breakpoint, +#endif /* !CONFIG_USER_ONLY */ +}; + +static void arm_v7m_class_init(ObjectClass *oc, void *data) +{ + ARMCPUClass *acc = ARM_CPU_CLASS(oc); + CPUClass *cc = CPU_CLASS(oc); + + acc->info = data; + cc->tcg_ops = &arm_v7m_tcg_ops; + cc->gdb_core_xml_file = "arm-m-profile.xml"; +} + +static const ARMCPUInfo arm_v7m_cpus[] = { + { .name = "cortex-m0", .initfn = cortex_m0_initfn, + .class_init = arm_v7m_class_init }, + { .name = "cortex-m3", .initfn = cortex_m3_initfn, + .class_init = arm_v7m_class_init }, + { .name = "cortex-m4", .initfn = cortex_m4_initfn, + .class_init = arm_v7m_class_init }, + { .name = "cortex-m7", .initfn = cortex_m7_initfn, + .class_init = arm_v7m_class_init }, + { .name = "cortex-m33", .initfn = cortex_m33_initfn, + .class_init = arm_v7m_class_init }, + { .name = "cortex-m55", .initfn = cortex_m55_initfn, + .class_init = arm_v7m_class_init }, +}; + +static void arm_v7m_cpu_register_types(void) +{ + size_t i; + + for (i = 0; i < ARRAY_SIZE(arm_v7m_cpus); ++i) { + arm_cpu_register(&arm_v7m_cpus[i]); + } +} + +type_init(arm_v7m_cpu_register_types) diff --git a/target/arm/tcg/cpu32.c b/target/arm/tcg/cpu32.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/cpu32.c +++ b/target/arm/tcg/cpu32.c @@ -XXX,XX +XXX,XX @@ #include "hw/boards.h" #endif #include "cpregs.h" -#if !defined(CONFIG_USER_ONLY) && defined(CONFIG_TCG) -#include "hw/intc/armv7m_nvic.h" -#endif /* Share AArch32 -cpu max features with AArch64. */ @@ -XXX,XX +XXX,XX @@ void aa32_max_features(ARMCPU *cpu) /* CPU models. These are not needed for the AArch64 linux-user build. */ #if !defined(CONFIG_USER_ONLY) || !defined(TARGET_AARCH64) -#if !defined(CONFIG_USER_ONLY) -static bool arm_v7m_cpu_exec_interrupt(CPUState *cs, int interrupt_request) -{ - CPUClass *cc = CPU_GET_CLASS(cs); - ARMCPU *cpu = ARM_CPU(cs); - CPUARMState *env = &cpu->env; - bool ret = false; - - /* - * ARMv7-M interrupt masking works differently than -A or -R. - * There is no FIQ/IRQ distinction. Instead of I and F bits - * masking FIQ and IRQ interrupts, an exception is taken only - * if it is higher priority than the current execution priority - * (which depends on state like BASEPRI, FAULTMASK and the - * currently active exception). - */ - if (interrupt_request & CPU_INTERRUPT_HARD - && (armv7m_nvic_can_take_pending_exception(env->nvic))) { - cs->exception_index = EXCP_IRQ; - cc->tcg_ops->do_interrupt(cs); - ret = true; - } - return ret; -} -#endif /* !CONFIG_USER_ONLY */ - static void arm926_initfn(Object *obj) { ARMCPU *cpu = ARM_CPU(obj); @@ -XXX,XX +XXX,XX @@ static void cortex_a15_initfn(Object *obj) define_arm_cp_regs(cpu, cortexa15_cp_reginfo); } -static void cortex_m0_initfn(Object *obj) -{ - ARMCPU *cpu = ARM_CPU(obj); - set_feature(&cpu->env, ARM_FEATURE_V6); - set_feature(&cpu->env, ARM_FEATURE_M); - - cpu->midr = 0x410cc200; - - /* - * These ID register values are not guest visible, because - * we do not implement the Main Extension. They must be set - * to values corresponding to the Cortex-M0's implemented - * features, because QEMU generally controls its emulation - * by looking at ID register fields. We use the same values as - * for the M3. - */ - cpu->isar.id_pfr0 = 0x00000030; - cpu->isar.id_pfr1 = 0x00000200; - cpu->isar.id_dfr0 = 0x00100000; - cpu->id_afr0 = 0x00000000; - cpu->isar.id_mmfr0 = 0x00000030; - cpu->isar.id_mmfr1 = 0x00000000; - cpu->isar.id_mmfr2 = 0x00000000; - cpu->isar.id_mmfr3 = 0x00000000; - cpu->isar.id_isar0 = 0x01141110; - cpu->isar.id_isar1 = 0x02111000; - cpu->isar.id_isar2 = 0x21112231; - cpu->isar.id_isar3 = 0x01111110; - cpu->isar.id_isar4 = 0x01310102; - cpu->isar.id_isar5 = 0x00000000; - cpu->isar.id_isar6 = 0x00000000; -} - -static void cortex_m3_initfn(Object *obj) -{ - ARMCPU *cpu = ARM_CPU(obj); - set_feature(&cpu->env, ARM_FEATURE_V7); - set_feature(&cpu->env, ARM_FEATURE_M); - set_feature(&cpu->env, ARM_FEATURE_M_MAIN); - cpu->midr = 0x410fc231; - cpu->pmsav7_dregion = 8; - cpu->isar.id_pfr0 = 0x00000030; - cpu->isar.id_pfr1 = 0x00000200; - cpu->isar.id_dfr0 = 0x00100000; - cpu->id_afr0 = 0x00000000; - cpu->isar.id_mmfr0 = 0x00000030; - cpu->isar.id_mmfr1 = 0x00000000; - cpu->isar.id_mmfr2 = 0x00000000; - cpu->isar.id_mmfr3 = 0x00000000; - cpu->isar.id_isar0 = 0x01141110; - cpu->isar.id_isar1 = 0x02111000; - cpu->isar.id_isar2 = 0x21112231; - cpu->isar.id_isar3 = 0x01111110; - cpu->isar.id_isar4 = 0x01310102; - cpu->isar.id_isar5 = 0x00000000; - cpu->isar.id_isar6 = 0x00000000; -} - -static void cortex_m4_initfn(Object *obj) -{ - ARMCPU *cpu = ARM_CPU(obj); - - set_feature(&cpu->env, ARM_FEATURE_V7); - set_feature(&cpu->env, ARM_FEATURE_M); - set_feature(&cpu->env, ARM_FEATURE_M_MAIN); - set_feature(&cpu->env, ARM_FEATURE_THUMB_DSP); - cpu->midr = 0x410fc240; /* r0p0 */ - cpu->pmsav7_dregion = 8; - cpu->isar.mvfr0 = 0x10110021; - cpu->isar.mvfr1 = 0x11000011; - cpu->isar.mvfr2 = 0x00000000; - cpu->isar.id_pfr0 = 0x00000030; - cpu->isar.id_pfr1 = 0x00000200; - cpu->isar.id_dfr0 = 0x00100000; - cpu->id_afr0 = 0x00000000; - cpu->isar.id_mmfr0 = 0x00000030; - cpu->isar.id_mmfr1 = 0x00000000; - cpu->isar.id_mmfr2 = 0x00000000; - cpu->isar.id_mmfr3 = 0x00000000; - cpu->isar.id_isar0 = 0x01141110; - cpu->isar.id_isar1 = 0x02111000; - cpu->isar.id_isar2 = 0x21112231; - cpu->isar.id_isar3 = 0x01111110; - cpu->isar.id_isar4 = 0x01310102; - cpu->isar.id_isar5 = 0x00000000; - cpu->isar.id_isar6 = 0x00000000; -} - -static void cortex_m7_initfn(Object *obj) -{ - ARMCPU *cpu = ARM_CPU(obj); - - set_feature(&cpu->env, ARM_FEATURE_V7); - set_feature(&cpu->env, ARM_FEATURE_M); - set_feature(&cpu->env, ARM_FEATURE_M_MAIN); - set_feature(&cpu->env, ARM_FEATURE_THUMB_DSP); - cpu->midr = 0x411fc272; /* r1p2 */ - cpu->pmsav7_dregion = 8; - cpu->isar.mvfr0 = 0x10110221; - cpu->isar.mvfr1 = 0x12000011; - cpu->isar.mvfr2 = 0x00000040; - cpu->isar.id_pfr0 = 0x00000030; - cpu->isar.id_pfr1 = 0x00000200; - cpu->isar.id_dfr0 = 0x00100000; - cpu->id_afr0 = 0x00000000; - cpu->isar.id_mmfr0 = 0x00100030; - cpu->isar.id_mmfr1 = 0x00000000; - cpu->isar.id_mmfr2 = 0x01000000; - cpu->isar.id_mmfr3 = 0x00000000; - cpu->isar.id_isar0 = 0x01101110; - cpu->isar.id_isar1 = 0x02112000; - cpu->isar.id_isar2 = 0x20232231; - cpu->isar.id_isar3 = 0x01111131; - cpu->isar.id_isar4 = 0x01310132; - cpu->isar.id_isar5 = 0x00000000; - cpu->isar.id_isar6 = 0x00000000; -} - -static void cortex_m33_initfn(Object *obj) -{ - ARMCPU *cpu = ARM_CPU(obj); - - set_feature(&cpu->env, ARM_FEATURE_V8); - set_feature(&cpu->env, ARM_FEATURE_M); - set_feature(&cpu->env, ARM_FEATURE_M_MAIN); - set_feature(&cpu->env, ARM_FEATURE_M_SECURITY); - set_feature(&cpu->env, ARM_FEATURE_THUMB_DSP); - cpu->midr = 0x410fd213; /* r0p3 */ - cpu->pmsav7_dregion = 16; - cpu->sau_sregion = 8; - cpu->isar.mvfr0 = 0x10110021; - cpu->isar.mvfr1 = 0x11000011; - cpu->isar.mvfr2 = 0x00000040; - cpu->isar.id_pfr0 = 0x00000030; - cpu->isar.id_pfr1 = 0x00000210; - cpu->isar.id_dfr0 = 0x00200000; - cpu->id_afr0 = 0x00000000; - cpu->isar.id_mmfr0 = 0x00101F40; - cpu->isar.id_mmfr1 = 0x00000000; - cpu->isar.id_mmfr2 = 0x01000000; - cpu->isar.id_mmfr3 = 0x00000000; - cpu->isar.id_isar0 = 0x01101110; - cpu->isar.id_isar1 = 0x02212000; - cpu->isar.id_isar2 = 0x20232232; - cpu->isar.id_isar3 = 0x01111131; - cpu->isar.id_isar4 = 0x01310132; - cpu->isar.id_isar5 = 0x00000000; - cpu->isar.id_isar6 = 0x00000000; - cpu->clidr = 0x00000000; - cpu->ctr = 0x8000c000; -} - -static void cortex_m55_initfn(Object *obj) -{ - ARMCPU *cpu = ARM_CPU(obj); - - set_feature(&cpu->env, ARM_FEATURE_V8); - set_feature(&cpu->env, ARM_FEATURE_V8_1M); - set_feature(&cpu->env, ARM_FEATURE_M); - set_feature(&cpu->env, ARM_FEATURE_M_MAIN); - set_feature(&cpu->env, ARM_FEATURE_M_SECURITY); - set_feature(&cpu->env, ARM_FEATURE_THUMB_DSP); - cpu->midr = 0x410fd221; /* r0p1 */ - cpu->revidr = 0; - cpu->pmsav7_dregion = 16; - cpu->sau_sregion = 8; - /* These are the MVFR* values for the FPU + full MVE configuration */ - cpu->isar.mvfr0 = 0x10110221; - cpu->isar.mvfr1 = 0x12100211; - cpu->isar.mvfr2 = 0x00000040; - cpu->isar.id_pfr0 = 0x20000030; - cpu->isar.id_pfr1 = 0x00000230; - cpu->isar.id_dfr0 = 0x10200000; - cpu->id_afr0 = 0x00000000; - cpu->isar.id_mmfr0 = 0x00111040; - cpu->isar.id_mmfr1 = 0x00000000; - cpu->isar.id_mmfr2 = 0x01000000; - cpu->isar.id_mmfr3 = 0x00000011; - cpu->isar.id_isar0 = 0x01103110; - cpu->isar.id_isar1 = 0x02212000; - cpu->isar.id_isar2 = 0x20232232; - cpu->isar.id_isar3 = 0x01111131; - cpu->isar.id_isar4 = 0x01310132; - cpu->isar.id_isar5 = 0x00000000; - cpu->isar.id_isar6 = 0x00000000; - cpu->clidr = 0x00000000; /* caches not implemented */ - cpu->ctr = 0x8303c003; -} - static const ARMCPRegInfo cortexr5_cp_reginfo[] = { /* Dummy the TCM region regs for the moment */ { .name = "ATCM", .cp = 15, .opc1 = 0, .crn = 9, .crm = 1, .opc2 = 0, @@ -XXX,XX +XXX,XX @@ static void pxa270c5_initfn(Object *obj) cpu->reset_sctlr = 0x00000078; } -static const TCGCPUOps arm_v7m_tcg_ops = { - .initialize = arm_translate_init, - .synchronize_from_tb = arm_cpu_synchronize_from_tb, - .debug_excp_handler = arm_debug_excp_handler, - .restore_state_to_opc = arm_restore_state_to_opc, - -#ifdef CONFIG_USER_ONLY - .record_sigsegv = arm_cpu_record_sigsegv, - .record_sigbus = arm_cpu_record_sigbus, -#else - .tlb_fill = arm_cpu_tlb_fill, - .cpu_exec_interrupt = arm_v7m_cpu_exec_interrupt, - .do_interrupt = arm_v7m_cpu_do_interrupt, - .do_transaction_failed = arm_cpu_do_transaction_failed, - .do_unaligned_access = arm_cpu_do_unaligned_access, - .adjust_watchpoint_address = arm_adjust_watchpoint_address, - .debug_check_watchpoint = arm_debug_check_watchpoint, - .debug_check_breakpoint = arm_debug_check_breakpoint, -#endif /* !CONFIG_USER_ONLY */ -}; - -static void arm_v7m_class_init(ObjectClass *oc, void *data) -{ - ARMCPUClass *acc = ARM_CPU_CLASS(oc); - CPUClass *cc = CPU_CLASS(oc); - - acc->info = data; - cc->tcg_ops = &arm_v7m_tcg_ops; - cc->gdb_core_xml_file = "arm-m-profile.xml"; -} - #ifndef TARGET_AARCH64 /* * -cpu max: a CPU with as many features enabled as our emulation supports. @@ -XXX,XX +XXX,XX @@ static const ARMCPUInfo arm_tcg_cpus[] = { { .name = "cortex-a8", .initfn = cortex_a8_initfn }, { .name = "cortex-a9", .initfn = cortex_a9_initfn }, { .name = "cortex-a15", .initfn = cortex_a15_initfn }, - { .name = "cortex-m0", .initfn = cortex_m0_initfn, - .class_init = arm_v7m_class_init }, - { .name = "cortex-m3", .initfn = cortex_m3_initfn, - .class_init = arm_v7m_class_init }, - { .name = "cortex-m4", .initfn = cortex_m4_initfn, - .class_init = arm_v7m_class_init }, - { .name = "cortex-m7", .initfn = cortex_m7_initfn, - .class_init = arm_v7m_class_init }, - { .name = "cortex-m33", .initfn = cortex_m33_initfn, - .class_init = arm_v7m_class_init }, - { .name = "cortex-m55", .initfn = cortex_m55_initfn, - .class_init = arm_v7m_class_init }, { .name = "cortex-r5", .initfn = cortex_r5_initfn }, { .name = "cortex-r5f", .initfn = cortex_r5f_initfn }, { .name = "cortex-r52", .initfn = cortex_r52_initfn }, diff --git a/target/arm/meson.build b/target/arm/meson.build index XXXXXXX..XXXXXXX 100644 --- a/target/arm/meson.build +++ b/target/arm/meson.build @@ -XXX,XX +XXX,XX @@ arm_system_ss.add(files( 'ptw.c', )) +arm_user_ss = ss.source_set() + subdir('hvf') if 'CONFIG_TCG' in config_all_accel @@ -XXX,XX +XXX,XX @@ endif target_arch += {'arm': arm_ss} target_system_arch += {'arm': arm_system_ss} +target_user_arch += {'arm': arm_user_ss} diff --git a/target/arm/tcg/meson.build b/target/arm/tcg/meson.build index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/meson.build +++ b/target/arm/tcg/meson.build @@ -XXX,XX +XXX,XX @@ arm_ss.add(when: 'TARGET_AARCH64', if_true: files( arm_system_ss.add(files( 'psci.c', )) + +arm_system_ss.add(when: 'CONFIG_ARM_V7M', if_true: files('cpu-v7m.c')) +arm_user_ss.add(when: 'TARGET_AARCH64', if_false: files('cpu-v7m.c')) -- 2.34.1