Re: [PATCH RFC 1/8] perf/core: Add *group_leader to perf_event_create_kernel_counter()

[Like Xu]

On 12/12/2022 9:23 pm, Marc Zyngier wrote:
> On Mon, 12 Dec 2022 12:58:37 +0000,
> Like Xu <like.xu.linux@xxxxxxxxx> wrote:
> > From: Like Xu <likexu@xxxxxxxxxxx>
> >
> > Like syscalls users, kernel-space perf_event creators may also use group
> > counters abstraction to gain pmu functionalities, and an in-kernel counter
> > groups behave much like normal 'single' counters, following the group
> > semantics-based behavior.
> >
> > No functional changes at this time. An example will be that KVM creates
> > Intel slot event as group leader and other topdown metric events to emulate
> > MSR_PERF_METRICS pmu capability for guests.
> >
> > Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> > Cc: Marc Zyngier <maz@xxxxxxxxxx>
> > Cc: Fenghua Yu <fenghua.yu@xxxxxxxxx>
> > Cc: kvmarm@xxxxxxxxxxxxxxx
> > Cc: linux-perf-users@xxxxxxxxxxxxxxx
> > Signed-off-by: Like Xu <likexu@xxxxxxxxxxx>
> > ---
> >   arch/arm64/kvm/pmu-emul.c                 | 4 ++--
> >   arch/x86/kernel/cpu/resctrl/pseudo_lock.c | 4 ++--
> >   arch/x86/kvm/pmu.c                        | 2 +-
> >   arch/x86/kvm/vmx/pmu_intel.c              | 2 +-
> >   include/linux/perf_event.h                | 1 +
> >   kernel/events/core.c                      | 4 +++-
> >   kernel/events/hw_breakpoint.c             | 4 ++--
> >   kernel/events/hw_breakpoint_test.c        | 2 +-
> >   kernel/watchdog_hld.c                     | 2 +-
> >   9 files changed, 14 insertions(+), 11 deletions(-)
> >
> > diff --git a/arch/arm64/kvm/pmu-emul.c b/arch/arm64/kvm/pmu-emul.c
> > index 24908400e190..11c3386bc86b 100644
> > --- a/arch/arm64/kvm/pmu-emul.c
> > +++ b/arch/arm64/kvm/pmu-emul.c
> > @@ -624,7 +624,7 @@ static void kvm_pmu_create_perf_event(struct kvm_pmc *pmc)
> >   
> >   	attr.sample_period = compute_period(pmc, kvm_pmu_get_pmc_value(pmc));
> >   
> > -	event = perf_event_create_kernel_counter(&attr, -1, current,
> > +	event = perf_event_create_kernel_counter(&attr, -1, current, NULL,
> >   						 kvm_pmu_perf_overflow, pmc);
>
> Wouldn't it be better to have a separate helper that takes the group leader
> as a parameter, and reimplement perf_event_create_kernel_counter() in term
> of this helper?
>
> 	M.

Applied. It makes changes more concise, thank you.