From patchwork Fri Jul 9 04:37:13 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Sergey Senozhatsky X-Patchwork-Id: 12366665 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-17.4 required=3.0 tests=BAYES_00,DKIMWL_WL_HIGH, DKIM_SIGNED,DKIM_VALID,HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id DA363C07E99 for ; Fri, 9 Jul 2021 04:39:48 +0000 (UTC) Received: from bombadil.infradead.org (bombadil.infradead.org [198.137.202.133]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id AAF6161435 for ; Fri, 9 Jul 2021 04:39:48 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org AAF6161435 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=chromium.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-arm-kernel-bounces+linux-arm-kernel=archiver.kernel.org@lists.infradead.org DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender: Content-Transfer-Encoding:Content-Type:List-Subscribe:List-Help:List-Post: List-Archive:List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To: Message-Id:Date:Subject:Cc:To:From:Reply-To:Content-ID:Content-Description: Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID: List-Owner; bh=w6qW6jOMCKLjx1DxL5sty16Mh6zBM2X//BHa7jlepvc=; b=4N8FFXnEAH/9Vy g7ILUnbEXc12idnLsbHytfow1IA80/iESDZlQGJXKFk/R26LRzC1gfP9TpkLoJPiWYsXtKsrDFglA +JkLJHqwiCu7H5wfFX6BrkXcxP8/HIfs/EQ8FpcRnDDL3OloNyON/e8eFCKAq1Fu5oLlNPOOxkNZq LH9A0mMnrTv+0jPFAq0HkOjlD5bCURtMfHtgde8a7wBM5/i8d03I84ZvSW8p3VOSbtm2e04FYp0Bd xBzT/g00Es+b8VGerw6aHXJHk+MJzW6j7nZOsKz0aD5DsPfvz0wiRt1KESmwTZZgh1alan8RBwzAC cIGatzXvWGM4tkc5Cv7g==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.94.2 #2 (Red Hat Linux)) id 1m1iHP-000kCl-HU; Fri, 09 Jul 2021 04:38:23 +0000 Received: from mail-pg1-x534.google.com ([2607:f8b0:4864:20::534]) by bombadil.infradead.org with esmtps (Exim 4.94.2 #2 (Red Hat Linux)) id 1m1iGm-000k5G-R7 for linux-arm-kernel@lists.infradead.org; Fri, 09 Jul 2021 04:37:46 +0000 Received: by mail-pg1-x534.google.com with SMTP id w15so8690639pgk.13 for ; Thu, 08 Jul 2021 21:37:44 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=chromium.org; s=google; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=ETT27YTe1Df1R34anpclNh/mV2vuSiYHLPmA2jkljAE=; b=dmlECuSGzQyT5BPgzoazb9gJYXKcWH37icvNOgZw3JI5O2XU1fCzPX6W+JxiavMd5Y scIVQeiaRcZ1l5KpD5zrGJ+/WAobQlvoW4/EiobILmMS+XW8TD2yOCS+Kijs4IGU+OiX HH1Dnh7r/BPjoIvuPN/qVdzlL2Go2M62+fnlo= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=ETT27YTe1Df1R34anpclNh/mV2vuSiYHLPmA2jkljAE=; b=e4v0ZVOu2VbvIxgJj+Q8W5PN0hORq/AXHyXviX1bZFa+R1OQnUuR7dgWWTnk9TNAiK Qp77qRAVXnFNnPZeWqAuq77hhBtloNOTsOVYVuJW9trQR8jlVu0ZGZlU4+uuTYxNSNyu slCpSENozTux8yJ/wY5rX+IckhCF1FaEfdOEF1DiclXxCX+ThW4rvEyCLKJhs4NNUhs+ BrXXeJ7MVD6N6IvuZoXepy6ndb9I24rDlj+3leujuHSoVGVyI1bKRCRSXy8Y3T38dIeE +V0X9x9VNyVCOafXAeg5g1xztnC7ScZfe2Skx7KiBa9bJBCK+/wYYpT/iHd5G8dw+1v/ SEdA== X-Gm-Message-State: AOAM532vVA1Qpr8qjP+hDxl8KW7psNYqYN3BLZixnNdpMTYhRy9/HYKU pOJfU0chAAm131/BlkLZtrCIfA== X-Google-Smtp-Source: ABdhPJyUpwfKVXdRnNKLdaV06jZZYOCtJWZE+pQ7Yig5Okf20u5ZB19Pmb0K+S0mixoNDsX0ipuwgg== X-Received: by 2002:a63:1226:: with SMTP id h38mr35402819pgl.376.1625805464184; Thu, 08 Jul 2021 21:37:44 -0700 (PDT) Received: from senozhatsky.flets-east.jp ([2409:10:2e40:5100:a30d:fc4:7834:ee08]) by smtp.gmail.com with ESMTPSA id y4sm4420760pfa.14.2021.07.08.21.37.41 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 08 Jul 2021 21:37:43 -0700 (PDT) From: Sergey Senozhatsky To: Marc Zyngier , Will Deacon Cc: Suleiman Souhlal , Joel Fernandes , linux-arm-kernel@lists.infradead.org, kvmarm@lists.cs.columbia.edu, linux-kernel@vger.kernel.org, virtualization@lists.linux-foundation.org, Sergey Senozhatsky Subject: [PATCHv2 4/4] arm64: add host pv-vcpu-state support Date: Fri, 9 Jul 2021 13:37:13 +0900 Message-Id: <20210709043713.887098-5-senozhatsky@chromium.org> X-Mailer: git-send-email 2.32.0.93.g670b81a890-goog In-Reply-To: <20210709043713.887098-1-senozhatsky@chromium.org> References: <20210709043713.887098-1-senozhatsky@chromium.org> MIME-Version: 1.0 X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20210708_213744_930487_B59B426F X-CRM114-Status: GOOD ( 23.65 ) X-BeenThere: linux-arm-kernel@lists.infradead.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+linux-arm-kernel=archiver.kernel.org@lists.infradead.org Add PV-vcpu-state support bits to the host. Host uses the guest PV-state per-CPU pointers to update the VCPU state each time it kvm_arch_vcpu_load() or kvm_arch_vcpu_put() the VCPU, so that guest scheduler can become aware of the fact that not all VCPUs are always available. Currently guest scheduler on amr64 always assumes that all CPUs are available because vcpu_is_preempted() is not implemented on arm64. - schbench -t 3 -m 3 -p 4096 Latency percentiles (usec) BASE ================================================ 50.0th: 1 (3556427 samples) 75.0th: 13 (879210 samples) 90.0th: 15 (893311 samples) 95.0th: 18 (159594 samples) *99.0th: 118 (224187 samples) 99.5th: 691 (28555 samples) 99.9th: 7384 (23076 samples) min=1, max=104218 avg worker transfer: 25192.00 ops/sec 98.41MB/s PATCHED ================================================ 50.0th: 1 (3507010 samples) 75.0th: 13 (1635775 samples) 90.0th: 16 (901271 samples) 95.0th: 24 (281051 samples) *99.0th: 114 (255581 samples) 99.5th: 382 (33051 samples) 99.9th: 6392 (26592 samples) min=1, max=83877 avg worker transfer: 28613.39 ops/sec 111.77MB/s - perf bench sched all ops/sec BASE PATCHED ================================================ 33452 36485 33541 39405 33365 36858 33455 38047 33449 37866 33616 34922 33479 34388 33594 37203 33458 35363 33704 35180 Student's T-test N Min Max Median Avg Stddev base 10 33365 33704 33479 33511.3 100.92467 patched 10 34388 39405 36858 36571.7 1607.454 Difference at 95.0% confidence 3060.4 +/- 1070.09 9.13244% +/- 3.19321% (Student's t, pooled s = 1138.88) Signed-off-by: Sergey Senozhatsky --- arch/arm64/include/asm/kvm_host.h | 23 +++++++++++ arch/arm64/kvm/Makefile | 3 +- arch/arm64/kvm/arm.c | 3 ++ arch/arm64/kvm/hypercalls.c | 11 ++++++ arch/arm64/kvm/pv-vcpu-state.c | 64 +++++++++++++++++++++++++++++++ 5 files changed, 103 insertions(+), 1 deletion(-) create mode 100644 arch/arm64/kvm/pv-vcpu-state.c diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h index 41911585ae0c..e782f4d0c916 100644 --- a/arch/arm64/include/asm/kvm_host.h +++ b/arch/arm64/include/asm/kvm_host.h @@ -381,6 +381,12 @@ struct kvm_vcpu_arch { u64 last_steal; gpa_t base; } steal; + + /* PV state of the VCPU */ + struct { + gpa_t base; + struct gfn_to_hva_cache ghc; + } vcpu_state; }; /* Pointer to the vcpu's SVE FFR for sve_{save,load}_state() */ @@ -695,6 +701,23 @@ static inline bool kvm_arm_is_pvtime_enabled(struct kvm_vcpu_arch *vcpu_arch) return (vcpu_arch->steal.base != GPA_INVALID); } +int kvm_init_vcpu_state(struct kvm_vcpu *vcpu, gfn_t addr); +int kvm_release_vcpu_state(struct kvm_vcpu *vcpu); + +static inline void kvm_arm_vcpu_state_init(struct kvm_vcpu_arch *vcpu_arch) +{ + vcpu_arch->vcpu_state.base = GPA_INVALID; + memset(&vcpu_arch->vcpu_state.ghc, 0, sizeof(struct gfn_to_hva_cache)); +} + +static inline bool +kvm_arm_is_vcpu_state_enabled(struct kvm_vcpu_arch *vcpu_arch) +{ + return (vcpu_arch->vcpu_state.base != GPA_INVALID); +} + +void kvm_update_vcpu_preempted(struct kvm_vcpu *vcpu, bool preempted); + void kvm_set_sei_esr(struct kvm_vcpu *vcpu, u64 syndrome); struct kvm_vcpu *kvm_mpidr_to_vcpu(struct kvm *kvm, unsigned long mpidr); diff --git a/arch/arm64/kvm/Makefile b/arch/arm64/kvm/Makefile index 989bb5dad2c8..2a3ee82c6d90 100644 --- a/arch/arm64/kvm/Makefile +++ b/arch/arm64/kvm/Makefile @@ -12,7 +12,8 @@ obj-$(CONFIG_KVM) += hyp/ kvm-y := $(KVM)/kvm_main.o $(KVM)/coalesced_mmio.o $(KVM)/eventfd.o \ $(KVM)/vfio.o $(KVM)/irqchip.o $(KVM)/binary_stats.o \ - arm.o mmu.o mmio.o psci.o perf.o hypercalls.o pvtime.o \ + arm.o mmu.o mmio.o psci.o perf.o hypercalls.o \ + pvtime.o pv-vcpu-state.o \ inject_fault.o va_layout.o handle_exit.o \ guest.o debug.o reset.o sys_regs.o \ vgic-sys-reg-v3.o fpsimd.o pmu.o \ diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c index e9a2b8f27792..43e995c9fddb 100644 --- a/arch/arm64/kvm/arm.c +++ b/arch/arm64/kvm/arm.c @@ -332,6 +332,7 @@ int kvm_arch_vcpu_create(struct kvm_vcpu *vcpu) kvm_arm_reset_debug_ptr(vcpu); kvm_arm_pvtime_vcpu_init(&vcpu->arch); + kvm_arm_vcpu_state_init(&vcpu->arch); vcpu->arch.hw_mmu = &vcpu->kvm->arch.mmu; @@ -429,10 +430,12 @@ void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu) if (vcpu_has_ptrauth(vcpu)) vcpu_ptrauth_disable(vcpu); kvm_arch_vcpu_load_debug_state_flags(vcpu); + kvm_update_vcpu_preempted(vcpu, false); } void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu) { + kvm_update_vcpu_preempted(vcpu, true); kvm_arch_vcpu_put_debug_state_flags(vcpu); kvm_arch_vcpu_put_fp(vcpu); if (has_vhe()) diff --git a/arch/arm64/kvm/hypercalls.c b/arch/arm64/kvm/hypercalls.c index 30da78f72b3b..95bcf86e0b6f 100644 --- a/arch/arm64/kvm/hypercalls.c +++ b/arch/arm64/kvm/hypercalls.c @@ -110,6 +110,9 @@ int kvm_hvc_call_handler(struct kvm_vcpu *vcpu) case ARM_SMCCC_HV_PV_TIME_FEATURES: val[0] = SMCCC_RET_SUCCESS; break; + case ARM_SMCCC_HV_PV_VCPU_STATE_FEATURES: + val[0] = SMCCC_RET_SUCCESS; + break; } break; case ARM_SMCCC_HV_PV_TIME_FEATURES: @@ -139,6 +142,14 @@ int kvm_hvc_call_handler(struct kvm_vcpu *vcpu) case ARM_SMCCC_TRNG_RND32: case ARM_SMCCC_TRNG_RND64: return kvm_trng_call(vcpu); + case ARM_SMCCC_HV_PV_VCPU_STATE_INIT: + if (kvm_init_vcpu_state(vcpu, smccc_get_arg1(vcpu)) == 0) + val[0] = SMCCC_RET_SUCCESS; + break; + case ARM_SMCCC_HV_PV_VCPU_STATE_RELEASE: + if (kvm_release_vcpu_state(vcpu) == 0) + val[0] = SMCCC_RET_SUCCESS; + break; default: return kvm_psci_call(vcpu); } diff --git a/arch/arm64/kvm/pv-vcpu-state.c b/arch/arm64/kvm/pv-vcpu-state.c new file mode 100644 index 000000000000..8496bb2a5966 --- /dev/null +++ b/arch/arm64/kvm/pv-vcpu-state.c @@ -0,0 +1,64 @@ +// SPDX-License-Identifier: GPL-2.0-only + +#include +#include + +#include +#include + +#include + +int kvm_init_vcpu_state(struct kvm_vcpu *vcpu, gpa_t addr) +{ + struct kvm *kvm = vcpu->kvm; + int ret; + u64 idx; + + if (kvm_arm_is_vcpu_state_enabled(&vcpu->arch)) + return 0; + + idx = srcu_read_lock(&kvm->srcu); + ret = kvm_gfn_to_hva_cache_init(vcpu->kvm, + &vcpu->arch.vcpu_state.ghc, + addr, + sizeof(struct vcpu_state)); + srcu_read_unlock(&kvm->srcu, idx); + + if (!ret) + vcpu->arch.vcpu_state.base = addr; + return ret; +} + +int kvm_release_vcpu_state(struct kvm_vcpu *vcpu) +{ + if (!kvm_arm_is_vcpu_state_enabled(&vcpu->arch)) + return 0; + + kvm_arm_vcpu_state_init(&vcpu->arch); + return 0; +} + +void kvm_update_vcpu_preempted(struct kvm_vcpu *vcpu, bool preempted) +{ + struct kvm *kvm = vcpu->kvm; + u64 idx; + + if (!kvm_arm_is_vcpu_state_enabled(&vcpu->arch)) + return; + + /* + * This function is called from atomic context, so we need to + * disable page faults. kvm_write_guest_cached() will call + * might_fault(). + */ + pagefault_disable(); + /* + * Need to take the SRCU lock because kvm_write_guest_offset_cached() + * calls kvm_memslots(); + */ + idx = srcu_read_lock(&kvm->srcu); + kvm_write_guest_cached(kvm, &vcpu->arch.vcpu_state.ghc, + &preempted, sizeof(bool)); + srcu_read_unlock(&kvm->srcu, idx); + pagefault_enable(); +}