From: "Mi, Dapeng" <dapeng1.mi@linux.intel.com>
To: Mingwei Zhang <mizhang@google.com>,
Sean Christopherson <seanjc@google.com>,
Paolo Bonzini <pbonzini@redhat.com>,
Xiong Zhang <xiong.y.zhang@intel.com>,
Kan Liang <kan.liang@intel.com>,
Zhenyu Wang <zhenyuw@linux.intel.com>,
Manali Shukla <manali.shukla@amd.com>,
Sandipan Das <sandipan.das@amd.com>
Cc: Jim Mattson <jmattson@google.com>,
Stephane Eranian <eranian@google.com>,
Ian Rogers <irogers@google.com>,
Namhyung Kim <namhyung@kernel.org>,
gce-passthrou-pmu-dev@google.com,
Samantha Alt <samantha.alt@intel.com>,
Zhiyuan Lv <zhiyuan.lv@intel.com>,
Yanfei Xu <yanfei.xu@intel.com>, maobibo <maobibo@loongson.cn>,
Like Xu <like.xu.linux@gmail.com>,
Peter Zijlstra <peterz@infradead.org>,
kvm@vger.kernel.org, linux-perf-users@vger.kernel.org
Subject: Re: [PATCH v2 30/54] KVM: x86/pmu: Implement the save/restore of PMU state for Intel CPU
Date: Tue, 14 May 2024 16:08:11 +0800 [thread overview]
Message-ID: <719d8331-331b-46b2-a2ff-fe5ff7fa4b5e@linux.intel.com> (raw)
In-Reply-To: <20240506053020.3911940-31-mizhang@google.com>
On 5/6/2024 1:29 PM, Mingwei Zhang wrote:
> Implement the save/restore of PMU state for pasthrough PMU in Intel. In
> passthrough mode, KVM owns exclusively the PMU HW when control flow goes to
> the scope of passthrough PMU. Thus, KVM needs to save the host PMU state
> and gains the full HW PMU ownership. On the contrary, host regains the
> ownership of PMU HW from KVM when control flow leaves the scope of
> passthrough PMU.
>
> Implement PMU context switches for Intel CPUs and opptunistically use
> rdpmcl() instead of rdmsrl() when reading counters since the former has
> lower latency in Intel CPUs.
It looks rdpmcl() optimization is removed from this patch, right? The
description is not identical with code.
>
> Co-developed-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
> Signed-off-by: Mingwei Zhang <mizhang@google.com>
The commit tags doesn't follow the rules in Linux document, we need to
change it in next version.
https://docs.kernel.org/process/submitting-patches.html#:~:text=Co%2Ddeveloped%2Dby%3A%20states,work%20on%20a%20single%20patch.
> ---
> arch/x86/kvm/pmu.c | 46 ++++++++++++++++++++++++++++++++++++
> arch/x86/kvm/vmx/pmu_intel.c | 41 +++++++++++++++++++++++++++++++-
> 2 files changed, 86 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/kvm/pmu.c b/arch/x86/kvm/pmu.c
> index 782b564bdf96..13197472e31d 100644
> --- a/arch/x86/kvm/pmu.c
> +++ b/arch/x86/kvm/pmu.c
> @@ -1068,14 +1068,60 @@ void kvm_pmu_passthrough_pmu_msrs(struct kvm_vcpu *vcpu)
>
> void kvm_pmu_save_pmu_context(struct kvm_vcpu *vcpu)
> {
> + struct kvm_pmu *pmu = vcpu_to_pmu(vcpu);
> + struct kvm_pmc *pmc;
> + u32 i;
> +
> lockdep_assert_irqs_disabled();
>
> static_call_cond(kvm_x86_pmu_save_pmu_context)(vcpu);
> +
> + /*
> + * Clear hardware selector MSR content and its counter to avoid
> + * leakage and also avoid this guest GP counter get accidentally
> + * enabled during host running when host enable global ctrl.
> + */
> + for (i = 0; i < pmu->nr_arch_gp_counters; i++) {
> + pmc = &pmu->gp_counters[i];
> + rdmsrl(pmc->msr_counter, pmc->counter);
I understood we want to use common code to manipulate PMU MSRs as much as
possible, but I don't think we should sacrifice performance. rdpmcl() has
better performance than rdmsrl(). If AMD CPUs doesn't support rdpmc
instruction, I think we should move this into vendor specific
xxx_save/restore_pmu_context helpers().
> + rdmsrl(pmc->msr_eventsel, pmc->eventsel);
> + if (pmc->counter)
> + wrmsrl(pmc->msr_counter, 0);
> + if (pmc->eventsel)
> + wrmsrl(pmc->msr_eventsel, 0);
> + }
> +
> + for (i = 0; i < pmu->nr_arch_fixed_counters; i++) {
> + pmc = &pmu->fixed_counters[i];
> + rdmsrl(pmc->msr_counter, pmc->counter);
> + if (pmc->counter)
> + wrmsrl(pmc->msr_counter, 0);
> + }
> }
>
> void kvm_pmu_restore_pmu_context(struct kvm_vcpu *vcpu)
> {
> + struct kvm_pmu *pmu = vcpu_to_pmu(vcpu);
> + struct kvm_pmc *pmc;
> + int i;
> +
> lockdep_assert_irqs_disabled();
>
> static_call_cond(kvm_x86_pmu_restore_pmu_context)(vcpu);
> +
> + /*
> + * No need to zero out unexposed GP/fixed counters/selectors since RDPMC
> + * in this case will be intercepted. Accessing to these counters and
> + * selectors will cause #GP in the guest.
> + */
> + for (i = 0; i < pmu->nr_arch_gp_counters; i++) {
> + pmc = &pmu->gp_counters[i];
> + wrmsrl(pmc->msr_counter, pmc->counter);
> + wrmsrl(pmc->msr_eventsel, pmu->gp_counters[i].eventsel);
> + }
> +
> + for (i = 0; i < pmu->nr_arch_fixed_counters; i++) {
> + pmc = &pmu->fixed_counters[i];
> + wrmsrl(pmc->msr_counter, pmc->counter);
> + }
> }
> diff --git a/arch/x86/kvm/vmx/pmu_intel.c b/arch/x86/kvm/vmx/pmu_intel.c
> index 7852ba25a240..a23cf9ca224e 100644
> --- a/arch/x86/kvm/vmx/pmu_intel.c
> +++ b/arch/x86/kvm/vmx/pmu_intel.c
> @@ -572,7 +572,7 @@ static void intel_pmu_refresh(struct kvm_vcpu *vcpu)
> }
>
> for (i = 0; i < pmu->nr_arch_fixed_counters; i++) {
> - pmu->fixed_counters[i].msr_eventsel = MSR_CORE_PERF_FIXED_CTR_CTRL;
> + pmu->fixed_counters[i].msr_eventsel = 0;
Why to initialize msr_eventsel to 0 instead of the real MSR address here?
> pmu->fixed_counters[i].msr_counter = MSR_CORE_PERF_FIXED_CTR0 + i;
> }
> }
> @@ -799,6 +799,43 @@ static void intel_passthrough_pmu_msrs(struct kvm_vcpu *vcpu)
> vmx_set_intercept_for_msr(vcpu, MSR_CORE_PERF_GLOBAL_OVF_CTRL, MSR_TYPE_RW, msr_intercept);
> }
>
> +static void intel_save_guest_pmu_context(struct kvm_vcpu *vcpu)
> +{
> + struct kvm_pmu *pmu = vcpu_to_pmu(vcpu);
> +
> + /* Global ctrl register is already saved at VM-exit. */
> + rdmsrl(MSR_CORE_PERF_GLOBAL_STATUS, pmu->global_status);
> + /* Clear hardware MSR_CORE_PERF_GLOBAL_STATUS MSR, if non-zero. */
> + if (pmu->global_status)
> + wrmsrl(MSR_CORE_PERF_GLOBAL_OVF_CTRL, pmu->global_status);
> +
> + rdmsrl(MSR_CORE_PERF_FIXED_CTR_CTRL, pmu->fixed_ctr_ctrl);
> + /*
> + * Clear hardware FIXED_CTR_CTRL MSR to avoid information leakage and
> + * also avoid these guest fixed counters get accidentially enabled
> + * during host running when host enable global ctrl.
> + */
> + if (pmu->fixed_ctr_ctrl)
> + wrmsrl(MSR_CORE_PERF_FIXED_CTR_CTRL, 0);
> +}
> +
> +static void intel_restore_guest_pmu_context(struct kvm_vcpu *vcpu)
> +{
> + struct kvm_pmu *pmu = vcpu_to_pmu(vcpu);
> + u64 global_status, toggle;
> +
> + /* Clear host global_ctrl MSR if non-zero. */
> + wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, 0);
> + rdmsrl(MSR_CORE_PERF_GLOBAL_STATUS, global_status);
> + toggle = pmu->global_status ^ global_status;
> + if (global_status & toggle)
> + wrmsrl(MSR_CORE_PERF_GLOBAL_OVF_CTRL, global_status & toggle);
> + if (pmu->global_status & toggle)
> + wrmsrl(MSR_CORE_PERF_GLOBAL_STATUS_SET, pmu->global_status & toggle);
> +
> + wrmsrl(MSR_CORE_PERF_FIXED_CTR_CTRL, pmu->fixed_ctr_ctrl);
> +}
> +
> struct kvm_pmu_ops intel_pmu_ops __initdata = {
> .rdpmc_ecx_to_pmc = intel_rdpmc_ecx_to_pmc,
> .msr_idx_to_pmc = intel_msr_idx_to_pmc,
> @@ -812,6 +849,8 @@ struct kvm_pmu_ops intel_pmu_ops __initdata = {
> .cleanup = intel_pmu_cleanup,
> .is_rdpmc_passthru_allowed = intel_is_rdpmc_passthru_allowed,
> .passthrough_pmu_msrs = intel_passthrough_pmu_msrs,
> + .save_pmu_context = intel_save_guest_pmu_context,
> + .restore_pmu_context = intel_restore_guest_pmu_context,
> .EVENTSEL_EVENT = ARCH_PERFMON_EVENTSEL_EVENT,
> .MAX_NR_GP_COUNTERS = KVM_INTEL_PMC_MAX_GENERIC,
> .MIN_NR_GP_COUNTERS = 1,
next prev parent reply other threads:[~2024-05-14 8:08 UTC|newest]
Thread overview: 93+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-05-06 5:29 [PATCH v2 00/54] Mediated Passthrough vPMU 2.0 for x86 Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 01/54] KVM: x86/pmu: Set enable bits for GP counters in PERF_GLOBAL_CTRL at "RESET" Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 02/54] KVM: x86: Snapshot if a vCPU's vendor model is AMD vs. Intel compatible Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 03/54] KVM: x86/pmu: Do not mask LVTPC when handling a PMI on AMD platforms Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 04/54] x86/msr: Define PerfCntrGlobalStatusSet register Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 05/54] x86/msr: Introduce MSR_CORE_PERF_GLOBAL_STATUS_SET Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 06/54] perf: Support get/put passthrough PMU interfaces Mingwei Zhang
2024-05-07 8:31 ` Peter Zijlstra
2024-05-08 4:13 ` Zhang, Xiong Y
2024-05-07 8:41 ` Peter Zijlstra
2024-05-08 4:54 ` Zhang, Xiong Y
2024-05-08 8:32 ` Peter Zijlstra
2024-05-06 5:29 ` [PATCH v2 07/54] perf: Add generic exclude_guest support Mingwei Zhang
2024-05-07 8:58 ` Peter Zijlstra
2024-05-06 5:29 ` [PATCH v2 08/54] perf/x86/intel: Support PERF_PMU_CAP_PASSTHROUGH_VPMU Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 09/54] perf: core/x86: Register a new vector for KVM GUEST PMI Mingwei Zhang
2024-05-07 9:12 ` Peter Zijlstra
2024-05-08 10:06 ` Yanfei Xu
2024-05-06 5:29 ` [PATCH v2 10/54] KVM: x86: Extract x86_set_kvm_irq_handler() function Mingwei Zhang
2024-05-07 9:18 ` Peter Zijlstra
2024-05-08 8:57 ` Zhang, Xiong Y
2024-05-06 5:29 ` [PATCH v2 11/54] KVM: x86/pmu: Register guest pmi handler for emulated PMU Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 12/54] perf: x86: Add x86 function to switch PMI handler Mingwei Zhang
2024-05-07 9:22 ` Peter Zijlstra
2024-05-08 6:58 ` Zhang, Xiong Y
2024-05-08 8:37 ` Peter Zijlstra
2024-05-09 7:30 ` Zhang, Xiong Y
2024-05-07 21:40 ` Chen, Zide
2024-05-08 3:44 ` Mi, Dapeng
2024-05-06 5:29 ` [PATCH v2 13/54] perf: core/x86: Forbid PMI handler when guest own PMU Mingwei Zhang
2024-05-07 9:33 ` Peter Zijlstra
2024-05-09 7:39 ` Zhang, Xiong Y
2024-05-06 5:29 ` [PATCH v2 14/54] perf: core/x86: Plumb passthrough PMU capability from x86_pmu to x86_pmu_cap Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 15/54] KVM: x86/pmu: Introduce enable_passthrough_pmu module parameter Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 16/54] KVM: x86/pmu: Plumb through pass-through PMU to vcpu for Intel CPUs Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 17/54] KVM: x86/pmu: Always set global enable bits in passthrough mode Mingwei Zhang
2024-05-08 4:18 ` Mi, Dapeng
2024-05-08 4:36 ` Mingwei Zhang
2024-05-08 6:27 ` Mi, Dapeng
2024-05-08 14:13 ` Sean Christopherson
2024-05-09 0:13 ` Mingwei Zhang
2024-05-09 0:30 ` Mi, Dapeng
2024-05-09 0:38 ` Mi, Dapeng
2024-05-06 5:29 ` [PATCH v2 18/54] KVM: x86/pmu: Add a helper to check if passthrough PMU is enabled Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 19/54] KVM: x86/pmu: Add host_perf_cap and initialize it in kvm_x86_vendor_init() Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 20/54] KVM: x86/pmu: Allow RDPMC pass through when all counters exposed to guest Mingwei Zhang
2024-05-08 21:55 ` Chen, Zide
2024-05-06 5:29 ` [PATCH v2 21/54] KVM: x86/pmu: Introduce macro PMU_CAP_PERF_METRICS Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 22/54] KVM: x86/pmu: Introduce PMU operator to check if rdpmc passthrough allowed Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 23/54] KVM: x86/pmu: Manage MSR interception for IA32_PERF_GLOBAL_CTRL Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 24/54] KVM: x86/pmu: Create a function prototype to disable MSR interception Mingwei Zhang
2024-05-08 22:03 ` Chen, Zide
2024-05-06 5:29 ` [PATCH v2 25/54] KVM: x86/pmu: Add intel_passthrough_pmu_msrs() to pass-through PMU MSRs Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 26/54] KVM: x86/pmu: Avoid legacy vPMU code when accessing global_ctrl in passthrough vPMU Mingwei Zhang
2024-05-08 21:48 ` Chen, Zide
2024-05-09 0:43 ` Mi, Dapeng
2024-05-09 1:29 ` Chen, Zide
2024-05-09 2:58 ` Mi, Dapeng
2024-05-06 5:29 ` [PATCH v2 27/54] KVM: x86/pmu: Exclude PMU MSRs in vmx_get_passthrough_msr_slot() Mingwei Zhang
2024-05-14 7:33 ` Mi, Dapeng
2024-05-06 5:29 ` [PATCH v2 28/54] KVM: x86/pmu: Add counter MSR and selector MSR index into struct kvm_pmc Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 29/54] KVM: x86/pmu: Introduce PMU operation prototypes for save/restore PMU context Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 30/54] KVM: x86/pmu: Implement the save/restore of PMU state for Intel CPU Mingwei Zhang
2024-05-14 8:08 ` Mi, Dapeng [this message]
2024-05-06 5:29 ` [PATCH v2 31/54] KVM: x86/pmu: Make check_pmu_event_filter() an exported function Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 32/54] KVM: x86/pmu: Allow writing to event selector for GP counters if event is allowed Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 33/54] KVM: x86/pmu: Allow writing to fixed counter selector if counter is exposed Mingwei Zhang
2024-05-06 5:29 ` [PATCH v2 34/54] KVM: x86/pmu: Switch IA32_PERF_GLOBAL_CTRL at VM boundary Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 35/54] KVM: x86/pmu: Exclude existing vLBR logic from the passthrough PMU Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 36/54] KVM: x86/pmu: Switch PMI handler at KVM context switch boundary Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 37/54] KVM: x86/pmu: Grab x86 core PMU for passthrough PMU VM Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 38/54] KVM: x86/pmu: Call perf_guest_enter() at PMU context switch Mingwei Zhang
2024-05-07 9:39 ` Peter Zijlstra
2024-05-08 4:22 ` Mi, Dapeng
2024-05-06 5:30 ` [PATCH v2 39/54] KVM: x86/pmu: Add support for PMU context switch at VM-exit/enter Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 40/54] KVM: x86/pmu: Introduce PMU operator to increment counter Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 41/54] KVM: x86/pmu: Introduce PMU operator for setting counter overflow Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 42/54] KVM: x86/pmu: Implement emulated counter increment for passthrough PMU Mingwei Zhang
2024-05-08 18:28 ` Chen, Zide
2024-05-09 1:11 ` Mi, Dapeng
2024-05-06 5:30 ` [PATCH v2 43/54] KVM: x86/pmu: Update pmc_{read,write}_counter() to disconnect perf API Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 44/54] KVM: x86/pmu: Disconnect counter reprogram logic from passthrough PMU Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 45/54] KVM: nVMX: Add nested virtualization support for " Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 46/54] perf/x86/amd/core: Set passthrough capability for host Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 47/54] KVM: x86/pmu/svm: Set passthrough capability for vcpus Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 48/54] KVM: x86/pmu/svm: Set enable_passthrough_pmu module parameter Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 49/54] KVM: x86/pmu/svm: Allow RDPMC pass through when all counters exposed to guest Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 50/54] KVM: x86/pmu/svm: Implement callback to disable MSR interception Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 51/54] KVM: x86/pmu/svm: Set GuestOnly bit and clear HostOnly bit when guest write to event selectors Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 52/54] KVM: x86/pmu/svm: Add registers to direct access list Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 53/54] KVM: x86/pmu/svm: Implement handlers to save and restore context Mingwei Zhang
2024-05-06 5:30 ` [PATCH v2 54/54] KVM: x86/pmu/svm: Wire up PMU filtering functionality for passthrough PMU Mingwei Zhang
2024-05-28 2:35 ` [PATCH v2 00/54] Mediated Passthrough vPMU 2.0 for x86 Ma, Yongwei
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=719d8331-331b-46b2-a2ff-fe5ff7fa4b5e@linux.intel.com \
--to=dapeng1.mi@linux.intel.com \
--cc=eranian@google.com \
--cc=gce-passthrou-pmu-dev@google.com \
--cc=irogers@google.com \
--cc=jmattson@google.com \
--cc=kan.liang@intel.com \
--cc=kvm@vger.kernel.org \
--cc=like.xu.linux@gmail.com \
--cc=linux-perf-users@vger.kernel.org \
--cc=manali.shukla@amd.com \
--cc=maobibo@loongson.cn \
--cc=mizhang@google.com \
--cc=namhyung@kernel.org \
--cc=pbonzini@redhat.com \
--cc=peterz@infradead.org \
--cc=samantha.alt@intel.com \
--cc=sandipan.das@amd.com \
--cc=seanjc@google.com \
--cc=xiong.y.zhang@intel.com \
--cc=yanfei.xu@intel.com \
--cc=zhenyuw@linux.intel.com \
--cc=zhiyuan.lv@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).