Re: [PATCH v4 4/6] kvm powerpc/book3s-apiv2: Introduce kvm-hv specific PMU

Athira Rajeev <[email protected]> Tue, 11 Mar 2025 15:23:55 +0530
Newsgroups org.kernel.vger.kvm-ppc,org.kernel.vger.kvm,org.ozlabs.lists.linuxppc-dev
Message-ID <[email protected]>

> On 10 Mar 2025, at 12:42=E2=80=AFPM, Vaibhav Jain =
<[email protected]> wrote:
>=20
> Athira Rajeev <[email protected]> writes:
>=20
>>> On 24 Feb 2025, at 6:45=E2=80=AFPM, Vaibhav Jain =
<[email protected]> wrote:
>>>=20
>>> Introduce a new PMU named 'kvm-hv' inside a new module named =
'kvm-hv-pmu'
>>> to report Book3s kvm-hv specific performance counters. This will =
expose
>>> KVM-HV specific performance attributes to user-space via kernel's =
PMU
>>> infrastructure and would enableusers to monitor active kvm-hv based =
guests.
>>>=20
>>> The patch creates necessary scaffolding to for the new PMU callbacks =
and
>>> introduces the new kernel module name 'kvm-hv-pmu' which is built =
with
>>> CONFIG_KVM_BOOK3S_HV_PMU. The patch doesn't introduce any =
perf-events yet,
>>> which will be introduced in later patches
>>>=20
>>> Signed-off-by: Vaibhav Jain <[email protected]>
>>>=20
>>> ---
>>> Changelog
>>>=20
>>> v3->v4:
>>> * Introduced a new kernel module named 'kmv-hv-pmu' to host the new =
PMU
>>> instead of building the as part of KVM-HV module. [ Maddy ]
>>> * Moved the code from arch/powerpc/kvm to arch/powerpc/perf [ =
Atheera ]
>>> * Added a new config named KVM_BOOK3S_HV_PMU to =
arch/powerpc/kvm/Kconfig
>>>=20
>>> v2->v3:
>>> * Fixed a build warning reported by kernel build robot.
>>> Link:
>>> =
https://lore.kernel.org/oe-kbuild-all/[email protected]
>>>=20
>>> v1->v2:
>>> * Fixed an issue of kvm-hv not loading on baremetal kvm [Gautam]
>>> ---
>>> arch/powerpc/kvm/Kconfig       |  13 ++++
>>> arch/powerpc/perf/Makefile     |   2 +
>>> arch/powerpc/perf/kvm-hv-pmu.c | 138 =
+++++++++++++++++++++++++++++++++
>>> 3 files changed, 153 insertions(+)
>>> create mode 100644 arch/powerpc/perf/kvm-hv-pmu.c
>>>=20
>>> diff --git a/arch/powerpc/kvm/Kconfig b/arch/powerpc/kvm/Kconfig
>>> index dbfdc126bf14..5f0ce19e7e27 100644
>>> --- a/arch/powerpc/kvm/Kconfig
>>> +++ b/arch/powerpc/kvm/Kconfig
>>> @@ -83,6 +83,7 @@ config KVM_BOOK3S_64_HV
>>> depends on KVM_BOOK3S_64 && PPC_POWERNV
>>> select KVM_BOOK3S_HV_POSSIBLE
>>> select KVM_GENERIC_MMU_NOTIFIER
>>> + select KVM_BOOK3S_HV_PMU
>>> select CMA
>>> help
>>> Support running unmodified book3s_64 guest kernels in
>>> @@ -171,6 +172,18 @@ config KVM_BOOK3S_HV_NESTED_PMU_WORKAROUND
>>> those buggy L1s which saves the L2 state, at the cost of performance
>>> in all nested-capable guest entry/exit.
>>>=20
>>> +config KVM_BOOK3S_HV_PMU
>>> + tristate "Hypervisor Perf events for KVM Book3s-HV"
>>> + depends on KVM_BOOK3S_64_HV && HV_PERF_CTRS
>>> + help
>>> +  Enable Book3s-HV Hypervisor Perf events PMU named 'kvm-hv'. These
>>> +  Perf events give an overview of hypervisor performance overall
>>> +  instead of a specific guests. Currently the PMU reports
>>> +  L0-Hypervisor stats on a kvm-hv enabled PSeries LPAR like:
>>> +  * Total/Used Guest-Heap
>>> +  * Total/Used Guest Page-table Memory
>>> +  * Total amount of Guest Page-table Memory reclaimed
>>> +
>>> config KVM_BOOKE_HV
>>> bool
>>>=20
>>> diff --git a/arch/powerpc/perf/Makefile b/arch/powerpc/perf/Makefile
>>> index ac2cf58d62db..7f53fcb7495a 100644
>>> --- a/arch/powerpc/perf/Makefile
>>> +++ b/arch/powerpc/perf/Makefile
>>> @@ -18,6 +18,8 @@ obj-$(CONFIG_HV_PERF_CTRS) +=3D hv-24x7.o =
hv-gpci.o hv-common.o
>>>=20
>>> obj-$(CONFIG_VPA_PMU) +=3D vpa-pmu.o
>>>=20
>>> +obj-$(CONFIG_KVM_BOOK3S_HV_PMU) +=3D kvm-hv-pmu.o
>>> +
>>> obj-$(CONFIG_PPC_8xx) +=3D 8xx-pmu.o
>>>=20
>>> obj-$(CONFIG_PPC64) +=3D $(obj64-y)
>>> diff --git a/arch/powerpc/perf/kvm-hv-pmu.c =
b/arch/powerpc/perf/kvm-hv-pmu.c
>>> new file mode 100644
>>> index 000000000000..c154f54e09e2
>>> --- /dev/null
>>> +++ b/arch/powerpc/perf/kvm-hv-pmu.c
>>> @@ -0,0 +1,138 @@
>>> +// SPDX-License-Identifier: GPL-2.0
>>> +/*
>>> + * Description: PMUs specific to running nested KVM-HV guests
>>> + * on Book3S processors (specifically POWER9 and later).
>>> + */
>>> +
>>> +#define pr_fmt(fmt)  "kvmppc-pmu: " fmt
>>> +
>>> +#include "asm-generic/local64.h"
>>> +#include <linux/kernel.h>
>>> +#include <linux/errno.h>
>>> +#include <linux/ratelimit.h>
>>> +#include <linux/kvm_host.h>
>>> +#include <linux/gfp_types.h>
>>> +#include <linux/pgtable.h>
>>> +#include <linux/perf_event.h>
>>> +#include <linux/spinlock_types.h>
>>> +#include <linux/spinlock.h>
>>> +
>>> +#include <asm/types.h>
>>> +#include <asm/kvm_ppc.h>
>>> +#include <asm/kvm_book3s.h>
>>> +#include <asm/mmu.h>
>>> +#include <asm/pgalloc.h>
>>> +#include <asm/pte-walk.h>
>>> +#include <asm/reg.h>
>>> +#include <asm/plpar_wrappers.h>
>>> +#include <asm/firmware.h>
>>> +
>>> +enum kvmppc_pmu_eventid {
>>> + KVMPPC_EVENT_MAX,
>>> +};
>>> +
>>> +static struct attribute *kvmppc_pmu_events_attr[] =3D {
>>> + NULL,
>>> +};
>>> +
>>> +static const struct attribute_group kvmppc_pmu_events_group =3D {
>>> + .name =3D "events",
>>> + .attrs =3D kvmppc_pmu_events_attr,
>>> +};
>>> +
>>> +PMU_FORMAT_ATTR(event, "config:0");
>>> +static struct attribute *kvmppc_pmu_format_attr[] =3D {
>>> + &format_attr_event.attr,
>>> + NULL,
>>> +};
>>> +
>>> +static struct attribute_group kvmppc_pmu_format_group =3D {
>>> + .name =3D "format",
>>> + .attrs =3D kvmppc_pmu_format_attr,
>>> +};
>>> +
>>> +static const struct attribute_group *kvmppc_pmu_attr_groups[] =3D {
>>> + &kvmppc_pmu_events_group,
>>> + &kvmppc_pmu_format_group,
>>> + NULL,
>>> +};
>>> +
>>> +static int kvmppc_pmu_event_init(struct perf_event *event)
>>> +{
>>> + unsigned int config =3D event->attr.config;
>>> +
>>> + pr_debug("%s: Event(%p) id=3D%llu cpu=3D%x on_cpu=3D%x config=3D%u",=

>>> + __func__, event, event->id, event->cpu,
>>> + event->oncpu, config);
>>> +
>>> + if (event->attr.type !=3D event->pmu->type)
>>> + return -ENOENT;
>>> +
>>> + if (config >=3D KVMPPC_EVENT_MAX)
>>> + return -EINVAL;
>>> +
>>> + local64_set(&event->hw.prev_count, 0);
>>> + local64_set(&event->count, 0);
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +static void kvmppc_pmu_del(struct perf_event *event, int flags)
>>> +{
>>> +}
>>> +
>>> +static int kvmppc_pmu_add(struct perf_event *event, int flags)
>>> +{
>>> + return 0;
>>> +}
>>> +
>>> +static void kvmppc_pmu_read(struct perf_event *event)
>>> +{
>>> +}
>>> +
>>> +/* L1 wide counters PMU */
>>> +static struct pmu kvmppc_pmu =3D {
>>> + .module =3D THIS_MODULE,
>>> + .task_ctx_nr =3D perf_sw_context,
>>> + .name =3D "kvm-hv",
>>> + .event_init =3D kvmppc_pmu_event_init,
>>> + .add =3D kvmppc_pmu_add,
>>> + .del =3D kvmppc_pmu_del,
>>> + .read =3D kvmppc_pmu_read,
>>> + .attr_groups =3D kvmppc_pmu_attr_groups,
>>> + .type =3D -1,
>>> +};
>>> +
>>> +static int __init kvmppc_register_pmu(void)
>>> +{
>>> + int rc =3D -EOPNOTSUPP;
>>> +
>>> + /* only support events for nestedv2 right now */
>>> + if (kvmhv_is_nestedv2()) {
>>=20
>> We don=E2=80=99t need PVR check here ? Description of module says =
this is
>> supported for power9 and later.
> The hcalls this module depends on, are only available to =
LPAR/KVM-Guest running with api-v2 support hence this is needed.
Ok, I understand we need kvmhv_is_nestedv2()
Doubt is whether we need PVR check here.

>=20
>>> + /* Setup done now register the PMU */
>>> + pr_info("Registering kvm-hv pmu");
>>> +
>>> + /* Register only if we arent already registered */
>> Not sure why we need this=E2=80=A6 Have you seen any issue without =
this ? I don=E2=80=99t see any similar check in =
arch/powerpc/perf/vpa-pmu.c ,
>>=20
> This check is taken from the previous version of this patch which
> prevented struct pmu initialization multiple times. However with
> now a seperate module this check is probably not needed.

Sure, Please check and remove if this is not needed.

Thanks
Athira
>=20
>>> + rc =3D (kvmppc_pmu.type =3D=3D -1) ?
>>> +     perf_pmu_register(&kvmppc_pmu, kvmppc_pmu.name,
>>> +       -1) : 0;
>>> + }
>>> +
>>> + return rc;
>>> +}
>>> +
>>> +static void __exit kvmppc_unregister_pmu(void)
>>> +{
>>> + if (kvmhv_is_nestedv2()) {
>>> + if (kvmppc_pmu.type !=3D -1)
>>> + perf_pmu_unregister(&kvmppc_pmu);
>>> +
>>> + pr_info("kvmhv_pmu unregistered.\n");
>>> + }
>>> +}
>>> +
>>> +module_init(kvmppc_register_pmu);
>>> +module_exit(kvmppc_unregister_pmu);
>>> +MODULE_DESCRIPTION("KVM PPC Book3s-hv PMU");
>>> +MODULE_AUTHOR("Vaibhav Jain <[email protected]>");
>>> +MODULE_LICENSE("GPL");
>>> --=20
>>> 2.48.1
>>>=20
>>>=20
>>>=20
>>=20
>=20
> --=20
> Cheers
> ~ Vaibhav