From: Colton Lewis <coltonlewis@google.com>
To: qemu-arm@nongnu.org, qemu-devel@nongnu.org, kvmarm@lists.linux.dev
Cc: Peter Maydell <peter.maydell@linaro.org>,
Paolo Bonzini <pbonzini@redhat.com>,
Marc Zyngier <maz@kernel.org>, Oliver Upton <oupton@kernel.org>,
Will Deacon <will@kernel.org>,
James Clark <james.clark@linaro.org>,
Mingwei Zhang <mizhang@google.com>,
kvm@vger.kernel.org, Colton Lewis <coltonlewis@google.com>
Subject: [PATCH] target/arm: Enable KVM PMU partitioning and counter limit
Date: Thu, 24 Sep 2026 17:30:51 +0000 [thread overview]
Message-ID: <20260924173051.2117637-1-coltonlewis@google.com> (raw)
In-Reply-To: <20260924172928.2110956-1-coltonlewis@google.com>
Add CPU properties to configure KVM ARM64 PMU counter partitioning:
- 'pmu-partition' (bool): enables KVM_ARM_VCPU_PMU_V3_ENABLE_PARTITION
- 'pmu-num-counters' (uint32): sets KVM_ARM_VCPU_PMU_V3_SET_NR_COUNTERS
When 'pmu-partition=on' and 'pmu-num-counters=N' are passed via -cpu,
QEMU enables hardware PMU partitioning and reserves N general-purpose
event counters for the guest prior to calling KVM_ARM_VCPU_PMU_V3_INIT.
Note that this patch is an example to demonstrate the KVM ARM64 PMU
Partitioning uAPI [1]. That is why it aborts immediately if setting the
device attributes does not succeed.
[1] https://lore.kernel.org/kvmarm/20260924172928.2110956-1-coltonlewis@google.com/
Signed-off-by: Colton Lewis <coltonlewis@google.com>
---
target/arm/cpu.c | 20 ++++++++++++++++++++
target/arm/cpu.h | 4 ++++
target/arm/kvm.c | 32 ++++++++++++++++++++++++++++++++
3 files changed, 56 insertions(+)
diff --git a/target/arm/cpu.c b/target/arm/cpu.c
index f58a1db843..7538c6a24c 100644
--- a/target/arm/cpu.c
+++ b/target/arm/cpu.c
@@ -1359,6 +1359,20 @@ static bool arm_get_pmu(Object *obj, Error **errp)
return cpu->has_pmu;
}
+static bool arm_get_pmu_partition(Object *obj, Error **errp)
+{
+ ARMCPU *cpu = ARM_CPU(obj);
+
+ return cpu->kvm_pmu_partition;
+}
+
+static void arm_set_pmu_partition(Object *obj, bool value, Error **errp)
+{
+ ARMCPU *cpu = ARM_CPU(obj);
+
+ cpu->kvm_pmu_partition = value;
+}
+
static void arm_set_pmu(Object *obj, bool value, Error **errp)
{
ARMCPU *cpu = ARM_CPU(obj);
@@ -1604,6 +1618,12 @@ static void arm_cpu_post_init(Object *obj)
if (arm_feature(&cpu->env, ARM_FEATURE_PMU)) {
cpu->has_pmu = true;
object_property_add_bool(obj, "pmu", arm_get_pmu, arm_set_pmu);
+ object_property_add_bool(obj, "pmu-partition",
+ arm_get_pmu_partition,
+ arm_set_pmu_partition);
+ object_property_add_uint32_ptr(obj, "pmu-num-counters",
+ &cpu->kvm_pmu_nr_counters,
+ OBJ_PROP_FLAG_READWRITE);
}
/*
diff --git a/target/arm/cpu.h b/target/arm/cpu.h
index e3f931dba2..1320e76fd4 100644
--- a/target/arm/cpu.h
+++ b/target/arm/cpu.h
@@ -1004,6 +1004,10 @@ struct ArchCPU {
bool has_el3;
/* CPU has PMU (Performance Monitor Unit) */
bool has_pmu;
+ /* CPU has PMU Partitioning enabled in KVM mode */
+ bool kvm_pmu_partition;
+ /* Number of PMU counters allocated to guest in KVM mode */
+ uint32_t kvm_pmu_nr_counters;
/* CPU has VFP */
bool has_vfp;
/* CPU has 32 VFP registers */
diff --git a/target/arm/kvm.c b/target/arm/kvm.c
index ed99be7fd8..3aa8243278 100644
--- a/target/arm/kvm.c
+++ b/target/arm/kvm.c
@@ -1846,6 +1846,10 @@ static bool kvm_arm_set_device_attr(ARMCPU *cpu, struct kvm_device_attr *attr,
return true;
}
+#ifndef KVM_ARM_VCPU_PMU_V3_ENABLE_PARTITION
+#define KVM_ARM_VCPU_PMU_V3_ENABLE_PARTITION 5
+#endif
+
void kvm_arm_pmu_init(ARMCPU *cpu)
{
struct kvm_device_attr attr = {
@@ -1856,6 +1860,34 @@ void kvm_arm_pmu_init(ARMCPU *cpu)
if (!cpu->has_pmu) {
return;
}
+
+ if (cpu->kvm_pmu_partition) {
+ uint32_t partition = 1;
+ struct kvm_device_attr part_attr = {
+ .group = KVM_ARM_VCPU_PMU_V3_CTRL,
+ .attr = KVM_ARM_VCPU_PMU_V3_ENABLE_PARTITION,
+ .addr = (intptr_t)&partition,
+ };
+
+ if (!kvm_arm_set_device_attr(cpu, &part_attr, "PMU partition")) {
+ error_report("failed to enable KVM PMU partition");
+ abort();
+ }
+ }
+
+ if (cpu->kvm_pmu_nr_counters > 0) {
+ struct kvm_device_attr cnt_attr = {
+ .group = KVM_ARM_VCPU_PMU_V3_CTRL,
+ .attr = KVM_ARM_VCPU_PMU_V3_SET_NR_COUNTERS,
+ .addr = (intptr_t)&cpu->kvm_pmu_nr_counters,
+ };
+
+ if (!kvm_arm_set_device_attr(cpu, &cnt_attr, "PMU nr_counters")) {
+ error_report("failed to set KVM PMU nr_counters");
+ abort();
+ }
+ }
+
if (!kvm_arm_set_device_attr(cpu, &attr, "PMU")) {
error_report("failed to init PMU");
abort();
base-commit: 014b1f730c0faa08067f8e9b5829e62ea3d4b177
--
2.56.0.rc1.315.gc6ed9934b7-goog
next prev parent reply other threads:[~2026-09-24 17:30 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-24 17:29 [PATCH v9 00/22] ARM64 PMU Partitioning Colton Lewis
2026-09-24 17:29 ` [PATCH v9 01/22] arm64: cpufeature: Add cpucap for HPMN0 Colton Lewis
2026-09-24 17:45 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 02/22] KVM: arm64: Reorganize PMU includes Colton Lewis
2026-09-24 17:38 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 03/22] KVM: arm64: Reorganize PMU functions Colton Lewis
2026-09-24 17:49 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 04/22] perf: arm_pmuv3: Generalize counter bitmasks Colton Lewis
2026-09-24 17:37 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 05/22] perf: arm_pmuv3: Move counter allocation mask to per-CPU struct pmu_hw_events Colton Lewis
2026-09-24 17:44 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 06/22] perf: arm_pmuv3: Check cntr_mask before using pmccntr Colton Lewis
2026-09-24 17:38 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 07/22] perf: arm_pmuv3: Allocate counter indices from high to low Colton Lewis
2026-09-24 17:37 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 08/22] KVM: arm64: Add initial scaffolding for Partitioned PMU Colton Lewis
2026-09-24 17:44 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 09/22] KVM: arm64: Set up FGT " Colton Lewis
2026-09-24 17:52 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 10/22] KVM: arm64: Add Partitioned PMU register trap handlers Colton Lewis
2026-09-24 17:50 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 11/22] KVM: arm64: Set up MDCR_EL2 to handle a Partitioned PMU Colton Lewis
2026-09-24 17:53 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 12/22] KVM: arm64: Context swap Partitioned PMU guest registers Colton Lewis
2026-09-24 17:55 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 13/22] KVM: arm64: Enforce PMU event filter at vcpu_load() Colton Lewis
2026-09-24 17:47 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 14/22] perf: Add perf_pmu_resched_update() Colton Lewis
2026-09-24 17:46 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 15/22] KVM: arm64: Allow kvm_vcpu_pmu_resync_el0() to resync filters in process context Colton Lewis
2026-09-24 17:51 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 16/22] KVM: arm64: Apply dynamic guest counter reservations Colton Lewis
2026-09-24 17:57 ` sashiko-bot
2026-09-30 15:28 ` James Clark
2026-10-01 21:33 ` Colton Lewis
2026-09-24 17:29 ` [PATCH v9 17/22] KVM: arm64: Implement lazy PMU context swaps Colton Lewis
2026-09-24 17:55 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 18/22] perf: arm_pmuv3: Handle IRQs for Partitioned PMU guest counters Colton Lewis
2026-09-24 18:03 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 19/22] KVM: arm64: Detect overflows for the Partitioned PMU Colton Lewis
2026-09-24 18:07 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 20/22] KVM: arm64: Add vCPU device attr to partition the PMU Colton Lewis
2026-09-24 17:56 ` sashiko-bot
2026-09-30 15:27 ` James Clark
2026-10-01 21:21 ` Colton Lewis
2026-09-24 17:29 ` [PATCH v9 21/22] KVM: selftests: Add find_bit to KVM library Colton Lewis
2026-09-24 17:52 ` sashiko-bot
2026-09-24 17:29 ` [PATCH v9 22/22] KVM: arm64: selftests: Add test case for Partitioned PMU Colton Lewis
2026-09-24 17:55 ` sashiko-bot
2026-09-24 17:30 ` Colton Lewis [this message]
2026-09-28 14:01 ` [PATCH] target/arm: Enable KVM PMU partitioning and counter limit Peter Maydell
2026-09-29 21:21 ` Colton Lewis
2026-09-30 11:35 ` Peter Maydell
2026-10-01 21:34 ` Colton Lewis
2026-09-30 15:25 ` [PATCH v9 00/22] ARM64 PMU Partitioning James Clark
2026-10-01 21:33 ` Colton Lewis
2026-09-30 15:26 ` James Clark
2026-10-01 21:33 ` Colton Lewis
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=20260924173051.2117637-1-coltonlewis@google.com \
--to=coltonlewis@google.com \
--cc=james.clark@linaro.org \
--cc=kvm@vger.kernel.org \
--cc=kvmarm@lists.linux.dev \
--cc=maz@kernel.org \
--cc=mizhang@google.com \
--cc=oupton@kernel.org \
--cc=pbonzini@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=qemu-arm@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=will@kernel.org \
/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