Received: by 2002:adf:db01:0:0:0:0:0 with SMTP id s1csp1250284wri; Thu, 31 Oct 2019 07:34:55 -0700 (PDT) X-Google-Smtp-Source: APXvYqyk4o8NI3wyiwtlLgxgBPVTGBkdeveuwuf32cWxSvFxeUFiLCQj6nNnO5H8fapo/AAoGADd X-Received: by 2002:ac8:2fba:: with SMTP id l55mr5981347qta.167.1572532495501; Thu, 31 Oct 2019 07:34:55 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1572532495; cv=none; d=google.com; s=arc-20160816; b=FmQijpygCQenbMHjqBx/kTApsrBTUinEvMTPdPu9vI45QTqn9htLjyVgPj6OYOuHWI kBYwyEehVzjH/ufPuI8Xms6WhdQEMyaUeM8fvR77T2XZBEqmTZltg9acThiJmiwPDwne s0m5apaEOdOtYsD99pHA2S/4xqB4LwTVVM14OpIgEg+t3XjRcUi8NLxzj1266Go1O0a2 GfTzwYp51I660wbZqeKG0KQrHHnPBhUtdh/MOkw5XpEK1uESKa7ZgSC7BQZSLbkIjiVU HXNGzwDgoTshvRORhmTo7WFNgRE00RIPwkkPcVUNb2RNjAMcFd8qQh0vK/SiAXAO/tXi MQrg== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:dkim-signature; bh=inFVYUzO9YS4A8Ly2zsnPPel9h9+jUPfLGKbk07PRRw=; b=ihheNl423LT7Yo9gjy6X4eTJFfnLaHrieCgxkDWM0igvHBLNdKzhVFqsVwoVvqhs4A 1VOafYRx2JNAp+U6i7utevEhyYWdXGM3wR8B+x7b2Fq/BKiAgqiut7NEQs8WIp+MHpYS A288InG0gDrHs1TI0vXMPL1UXhdW10sYNYj13cktCZjI4EZ1fKsaxkMPaZiAyC93xNsF 8NCaCv+//w/avCCHamlYI+B0Y9LMMY/dXXqI4LcFx1+1PObSVjd/5zEwXmPUwXqGFkQC 6BEiXiH2LPsU/4JJxyxE1NNH7CQH1zxhJGLmHGYkCjd67Xav8vNkd2jaqHYR4w2qRVKv Qqwg== ARC-Authentication-Results: i=1; mx.google.com; dkim=pass header.i=@redhat.com header.s=mimecast20190719 header.b="Un/THii0"; spf=pass (google.com: domain of drjones@redhat.com designates 205.139.110.61 as permitted sender) smtp.mailfrom=drjones@redhat.com; dmarc=pass (p=NONE sp=NONE dis=NONE) header.from=redhat.com Return-Path: Received: from us-smtp-delivery-1.mimecast.com (us-smtp-2.mimecast.com. [205.139.110.61]) by mx.google.com with ESMTPS id w11si3495753qvz.128.2019.10.31.07.34.55 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Thu, 31 Oct 2019 07:34:55 -0700 (PDT) Received-SPF: pass (google.com: domain of drjones@redhat.com designates 205.139.110.61 as permitted sender) client-ip=205.139.110.61; Authentication-Results: mx.google.com; dkim=pass header.i=@redhat.com header.s=mimecast20190719 header.b="Un/THii0"; spf=pass (google.com: domain of drjones@redhat.com designates 205.139.110.61 as permitted sender) smtp.mailfrom=drjones@redhat.com; dmarc=pass (p=NONE sp=NONE dis=NONE) header.from=redhat.com DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1572532495; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=inFVYUzO9YS4A8Ly2zsnPPel9h9+jUPfLGKbk07PRRw=; b=Un/THii0LKlktL/M8zgn7rddloKEbRk1TurBbbDgrF2zn3hUtWamdsX0NkwW+eSVVlHaa4 tA7ZprfZoXMYIR2XwtEVdmoga25AAgNznD/y28o0YK01MFKSUGFnMwkiwLswxPGWoAvosH FmGY9l5ZawRe3pFcSV8GrMJylov0ZCs= Received: from mimecast-mx01.redhat.com (mimecast-mx01.redhat.com [209.132.183.4]) (Using TLS) by relay.mimecast.com with ESMTP id us-mta-364--nDddkYRONChI-JDj1g9jg-1; Thu, 31 Oct 2019 10:34:46 -0400 Received: from smtp.corp.redhat.com (int-mx03.intmail.prod.int.phx2.redhat.com [10.5.11.13]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mimecast-mx01.redhat.com (Postfix) with ESMTPS id 06C101800D55; Thu, 31 Oct 2019 14:34:45 +0000 (UTC) Received: from kamzik.brq.redhat.com (ovpn-117-53.ams2.redhat.com [10.36.117.53]) by smtp.corp.redhat.com (Postfix) with ESMTP id 4758C60852; Thu, 31 Oct 2019 14:34:17 +0000 (UTC) From: Andrew Jones To: qemu-devel@nongnu.org, qemu-arm@nongnu.org, peter.maydell@linaro.org Cc: richard.henderson@linaro.org, eric.auger@redhat.com, beata.michalska@linaro.org, m.mizuma@jp.fujitsu.com, alex.bennee@linaro.org, Dave.Martin@arm.com, armbru@redhat.com, eblake@redhat.com, imammedo@redhat.com Subject: [PATCH v8 5/9] target/arm/kvm64: Add kvm_arch_get/put_sve Date: Thu, 31 Oct 2019 15:27:30 +0100 Message-Id: <20191031142734.8590-6-drjones@redhat.com> In-Reply-To: <20191031142734.8590-1-drjones@redhat.com> References: <20191031142734.8590-1-drjones@redhat.com> MIME-Version: 1.0 X-Scanned-By: MIMEDefang 2.79 on 10.5.11.13 X-MC-Unique: -nDddkYRONChI-JDj1g9jg-1 X-Mimecast-Spam-Score: 0 Content-Type: text/plain; charset=WINDOWS-1252 Content-Transfer-Encoding: quoted-printable X-TUID: 8M8Mz3BnP/RO These are the SVE equivalents to kvm_arch_get/put_fpsimd. Note, the swabbing is different than it is for fpsmid because the vector format is a little-endian stream of words. Signed-off-by: Andrew Jones Reviewed-by: Richard Henderson Reviewed-by: Eric Auger Tested-by: Masayoshi Mizuma --- target/arm/kvm64.c | 183 ++++++++++++++++++++++++++++++++++++++------- 1 file changed, 155 insertions(+), 28 deletions(-) diff --git a/target/arm/kvm64.c b/target/arm/kvm64.c index 28f6db57d5ee..4c0b11d105a4 100644 --- a/target/arm/kvm64.c +++ b/target/arm/kvm64.c @@ -671,11 +671,12 @@ int kvm_arch_destroy_vcpu(CPUState *cs) bool kvm_arm_reg_syncs_via_cpreg_list(uint64_t regidx) { /* Return true if the regidx is a register we should synchronize - * via the cpreg_tuples array (ie is not a core reg we sync by - * hand in kvm_arch_get/put_registers()) + * via the cpreg_tuples array (ie is not a core or sve reg that + * we sync by hand in kvm_arch_get/put_registers()) */ switch (regidx & KVM_REG_ARM_COPROC_MASK) { case KVM_REG_ARM_CORE: + case KVM_REG_ARM64_SVE: return false; default: return true; @@ -721,10 +722,8 @@ int kvm_arm_cpreg_level(uint64_t regidx) =20 static int kvm_arch_put_fpsimd(CPUState *cs) { - ARMCPU *cpu =3D ARM_CPU(cs); - CPUARMState *env =3D &cpu->env; + CPUARMState *env =3D &ARM_CPU(cs)->env; struct kvm_one_reg reg; - uint32_t fpr; int i, ret; =20 for (i =3D 0; i < 32; i++) { @@ -742,17 +741,73 @@ static int kvm_arch_put_fpsimd(CPUState *cs) } } =20 - reg.addr =3D (uintptr_t)(&fpr); - fpr =3D vfp_get_fpsr(env); - reg.id =3D AARCH64_SIMD_CTRL_REG(fp_regs.fpsr); - ret =3D kvm_vcpu_ioctl(cs, KVM_SET_ONE_REG, ®); - if (ret) { - return ret; + return 0; +} + +/* + * SVE registers are encoded in KVM's memory in an endianness-invariant fo= rmat. + * The byte at offset i from the start of the in-memory representation con= tains + * the bits [(7 + 8 * i) : (8 * i)] of the register value. As this means t= he + * lowest offsets are stored in the lowest memory addresses, then that nea= rly + * matches QEMU's representation, which is to use an array of host-endian + * uint64_t's, where the lower offsets are at the lower indices. To comple= te + * the translation we just need to byte swap the uint64_t's on big-endian = hosts. + */ +static uint64_t *sve_bswap64(uint64_t *dst, uint64_t *src, int nr) +{ +#ifdef HOST_WORDS_BIGENDIAN + int i; + + for (i =3D 0; i < nr; ++i) { + dst[i] =3D bswap64(src[i]); } =20 - reg.addr =3D (uintptr_t)(&fpr); - fpr =3D vfp_get_fpcr(env); - reg.id =3D AARCH64_SIMD_CTRL_REG(fp_regs.fpcr); + return dst; +#else + return src; +#endif +} + +/* + * KVM SVE registers come in slices where ZREGs have a slice size of 2048 = bits + * and PREGS and the FFR have a slice size of 256 bits. However we simply = hard + * code the slice index to zero for now as it's unlikely we'll need more t= han + * one slice for quite some time. + */ +static int kvm_arch_put_sve(CPUState *cs) +{ + ARMCPU *cpu =3D ARM_CPU(cs); + CPUARMState *env =3D &cpu->env; + uint64_t tmp[ARM_MAX_VQ * 2]; + uint64_t *r; + struct kvm_one_reg reg; + int n, ret; + + for (n =3D 0; n < KVM_ARM64_SVE_NUM_ZREGS; ++n) { + r =3D sve_bswap64(tmp, &env->vfp.zregs[n].d[0], cpu->sve_max_vq * = 2); + reg.addr =3D (uintptr_t)r; + reg.id =3D KVM_REG_ARM64_SVE_ZREG(n, 0); + ret =3D kvm_vcpu_ioctl(cs, KVM_SET_ONE_REG, ®); + if (ret) { + return ret; + } + } + + for (n =3D 0; n < KVM_ARM64_SVE_NUM_PREGS; ++n) { + r =3D sve_bswap64(tmp, r =3D &env->vfp.pregs[n].p[0], + DIV_ROUND_UP(cpu->sve_max_vq * 2, 8)); + reg.addr =3D (uintptr_t)r; + reg.id =3D KVM_REG_ARM64_SVE_PREG(n, 0); + ret =3D kvm_vcpu_ioctl(cs, KVM_SET_ONE_REG, ®); + if (ret) { + return ret; + } + } + + r =3D sve_bswap64(tmp, &env->vfp.pregs[FFR_PRED_NUM].p[0], + DIV_ROUND_UP(cpu->sve_max_vq * 2, 8)); + reg.addr =3D (uintptr_t)r; + reg.id =3D KVM_REG_ARM64_SVE_FFR(0); ret =3D kvm_vcpu_ioctl(cs, KVM_SET_ONE_REG, ®); if (ret) { return ret; @@ -765,6 +820,7 @@ int kvm_arch_put_registers(CPUState *cs, int level) { struct kvm_one_reg reg; uint64_t val; + uint32_t fpr; int i, ret; unsigned int el; =20 @@ -855,7 +911,27 @@ int kvm_arch_put_registers(CPUState *cs, int level) } } =20 - ret =3D kvm_arch_put_fpsimd(cs); + if (cpu_isar_feature(aa64_sve, cpu)) { + ret =3D kvm_arch_put_sve(cs); + } else { + ret =3D kvm_arch_put_fpsimd(cs); + } + if (ret) { + return ret; + } + + reg.addr =3D (uintptr_t)(&fpr); + fpr =3D vfp_get_fpsr(env); + reg.id =3D AARCH64_SIMD_CTRL_REG(fp_regs.fpsr); + ret =3D kvm_vcpu_ioctl(cs, KVM_SET_ONE_REG, ®); + if (ret) { + return ret; + } + + reg.addr =3D (uintptr_t)(&fpr); + fpr =3D vfp_get_fpcr(env); + reg.id =3D AARCH64_SIMD_CTRL_REG(fp_regs.fpcr); + ret =3D kvm_vcpu_ioctl(cs, KVM_SET_ONE_REG, ®); if (ret) { return ret; } @@ -878,10 +954,8 @@ int kvm_arch_put_registers(CPUState *cs, int level) =20 static int kvm_arch_get_fpsimd(CPUState *cs) { - ARMCPU *cpu =3D ARM_CPU(cs); - CPUARMState *env =3D &cpu->env; + CPUARMState *env =3D &ARM_CPU(cs)->env; struct kvm_one_reg reg; - uint32_t fpr; int i, ret; =20 for (i =3D 0; i < 32; i++) { @@ -899,21 +973,53 @@ static int kvm_arch_get_fpsimd(CPUState *cs) } } =20 - reg.addr =3D (uintptr_t)(&fpr); - reg.id =3D AARCH64_SIMD_CTRL_REG(fp_regs.fpsr); - ret =3D kvm_vcpu_ioctl(cs, KVM_GET_ONE_REG, ®); - if (ret) { - return ret; + return 0; +} + +/* + * KVM SVE registers come in slices where ZREGs have a slice size of 2048 = bits + * and PREGS and the FFR have a slice size of 256 bits. However we simply = hard + * code the slice index to zero for now as it's unlikely we'll need more t= han + * one slice for quite some time. + */ +static int kvm_arch_get_sve(CPUState *cs) +{ + ARMCPU *cpu =3D ARM_CPU(cs); + CPUARMState *env =3D &cpu->env; + struct kvm_one_reg reg; + uint64_t *r; + int n, ret; + + for (n =3D 0; n < KVM_ARM64_SVE_NUM_ZREGS; ++n) { + r =3D &env->vfp.zregs[n].d[0]; + reg.addr =3D (uintptr_t)r; + reg.id =3D KVM_REG_ARM64_SVE_ZREG(n, 0); + ret =3D kvm_vcpu_ioctl(cs, KVM_GET_ONE_REG, ®); + if (ret) { + return ret; + } + sve_bswap64(r, r, cpu->sve_max_vq * 2); } - vfp_set_fpsr(env, fpr); =20 - reg.addr =3D (uintptr_t)(&fpr); - reg.id =3D AARCH64_SIMD_CTRL_REG(fp_regs.fpcr); + for (n =3D 0; n < KVM_ARM64_SVE_NUM_PREGS; ++n) { + r =3D &env->vfp.pregs[n].p[0]; + reg.addr =3D (uintptr_t)r; + reg.id =3D KVM_REG_ARM64_SVE_PREG(n, 0); + ret =3D kvm_vcpu_ioctl(cs, KVM_GET_ONE_REG, ®); + if (ret) { + return ret; + } + sve_bswap64(r, r, DIV_ROUND_UP(cpu->sve_max_vq * 2, 8)); + } + + r =3D &env->vfp.pregs[FFR_PRED_NUM].p[0]; + reg.addr =3D (uintptr_t)r; + reg.id =3D KVM_REG_ARM64_SVE_FFR(0); ret =3D kvm_vcpu_ioctl(cs, KVM_GET_ONE_REG, ®); if (ret) { return ret; } - vfp_set_fpcr(env, fpr); + sve_bswap64(r, r, DIV_ROUND_UP(cpu->sve_max_vq * 2, 8)); =20 return 0; } @@ -923,6 +1029,7 @@ int kvm_arch_get_registers(CPUState *cs) struct kvm_one_reg reg; uint64_t val; unsigned int el; + uint32_t fpr; int i, ret; =20 ARMCPU *cpu =3D ARM_CPU(cs); @@ -1012,10 +1119,30 @@ int kvm_arch_get_registers(CPUState *cs) env->spsr =3D env->banked_spsr[i]; } =20 - ret =3D kvm_arch_get_fpsimd(cs); + if (cpu_isar_feature(aa64_sve, cpu)) { + ret =3D kvm_arch_get_sve(cs); + } else { + ret =3D kvm_arch_get_fpsimd(cs); + } + if (ret) { + return ret; + } + + reg.addr =3D (uintptr_t)(&fpr); + reg.id =3D AARCH64_SIMD_CTRL_REG(fp_regs.fpsr); + ret =3D kvm_vcpu_ioctl(cs, KVM_GET_ONE_REG, ®); + if (ret) { + return ret; + } + vfp_set_fpsr(env, fpr); + + reg.addr =3D (uintptr_t)(&fpr); + reg.id =3D AARCH64_SIMD_CTRL_REG(fp_regs.fpcr); + ret =3D kvm_vcpu_ioctl(cs, KVM_GET_ONE_REG, ®); if (ret) { return ret; } + vfp_set_fpcr(env, fpr); =20 ret =3D kvm_get_vcpu_events(cpu); if (ret) { --=20 2.21.0